US 10,754,309 B2 · Auto defect screening using adaptive machine learning in semiconductor device manufacturing flow

Abstract

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.

Claims12

Dependent claims

A method for defect screening in semiconductor manufacturing, comprising:

inspecting a die to collect defect candidate data from an image of the die, the defect candidate data comprising design data;
validating the defect candidate data to form defect of interest data comprising the defect candidate data, the design data, and a label indicating that the defect candidate data is one of a real defect or a nuisance;
updating a critical signature data library comprising a plurality of defect of interest data elements to comprise the defect of interest data;
compiling model training data comprising the plurality of defect of interest data elements from the critical signature library;
training a machine learning model with the model training data to generate a trained machine learning model;
collecting, using an optical sensor, an image of a second die;
inspecting the image of the second die with the trained machine learning model, the trained machine learning model:
identifying a second defect candidate data from the image of the second die, the second defect data comprising second design data;
and
identifying the second defect candidate data as a defect;
and
identifying the second die as containing a defect.

9 more elements

A non-transitory computer readable medium comprising instructions that, when executed by a processor of a processing system, cause the processing system to perform a method for improving semiconductor device fabrication, the method comprising:

inspecting a die to collect defect candidate data from an image of the die, the defect candidate data comprising design data;
validating the defect candidate data to form defect of interest data comprising the defect candidate data, the design data, and a label indicating that the defect candidate data is one of a real defect or a nuisance;
updating a critical signature data library comprising a plurality of defect of interest data elements to comprise the defect of interest data;
compiling model training data comprising the plurality of defect of interest data elements from the critical signature library;
training a machine learning model with the model training data to generate a trained machine learning model;
collecting, using an optical sensor, an image of a second die;
inspecting the image of the second die with the trained machine learning model, the trained machine learning model:
identifying a second defect candidate data from the image of the second die, the second defect data comprising second design data;
and
identifying the second defect candidate data as a defect;
and
identifying the second die as containing a defect.

9 more elements

A system for improving semiconductor device fabrication, comprising:

a memory comprising computer-executable instructions;
a processor configured to execute the computer-executable instructions and cause the processing system to perform a method for improving semiconductor device fabrication, the method comprising:
inspecting a die to collect defect candidate data from an image of the die, the defect candidate data comprising design data;
validating the defect candidate data to form defect of interest data comprising the defect candidate data, the design data, and a label indicating that the defect candidate data is one of a real defect or a nuisance;
updating a critical signature data library comprising a plurality of defect of interest data elements to comprise the defect of interest data;
compiling model training data comprising the plurality of defect of interest data elements from the critical signature library;
training a machine learning model with the model training data to generate a trained machine learning model;
collecting, using an optical sensor, an image of a second die;
inspecting the image of the second die with the trained machine learning model, the trained machine learning model:
identifying a second defect candidate data from the image of the second die, the second defect data comprising second design data;
and
identifying the second defect candidate data as a defect;
and
identifying the second die as containing a defect.

11 more elements

Dependencies are inferred from the claim text.

Auto defect screening using adaptive machine learning in semiconductor device manufacturing flow

Life of this patent

Auto defect screening using adaptive machine learning in semiconductor device manufacturing flow

Life of this patent

Abstract

Claims12