Yield estimation and control

Abstract

A defect prediction method for a device manufacturing process involving production substrates processed by a lithographic apparatus, the method including training a classification model using a training set including measured or determined values of a process parameter associated with the production substrates processed by the device manufacturing process and an indication regarding existence of defects associated with the production substrates processed in the device manufacturing process under the values of the process parameter, and producing an output from the classification model that indicates a prediction of a defect for a substrate.

What is claimed is: 
     
         1 . A computer-implemented defect prediction method for a device manufacturing process involving production substrates processed by a lithographic apparatus, the method comprising:
 training a classification model using a training set comprising measured or determined values of a process parameter associated with the production substrates processed by the device manufacturing process and an indication regarding existence of defects associated with the production substrates processed in the device manufacturing process under the values of the process parameter; and   producing an output from the classification model that indicates a prediction of a defect for a substrate.   
     
     
         2 . The method of  claim 1 , comprising training the classification model using a further training set comprising further measured or determined values of a process parameter associated with production substrates processed by the device manufacturing process and an indication regarding existence of defects associated with the production substrates processed in the device manufacturing process under the further values of the process parameter. 
     
     
         3 . The method of  claim 2 , wherein at least some of the further values are generated after training the classification model using the measured or determined values. 
     
     
         4 . The method of  claim 2 , wherein the further training set comprises at least a portion of the measured or determined values in addition to the further values. 
     
     
         5 . The method of  claim 1 , further comprising repeatedly performing the training based on further measured or determined values of the process parameter associated with further production substrates processed by the device manufacturing process. 
     
     
         6 . The method of  claim 1 , further comprising calculating a probability of the defect for the substrate using the classification model. 
     
     
         7 . The method of  claim 6 , further comprising adjusting a parameter of the device manufacturing process, a parameter of a layout to be patterned onto a substrate, or both, using the probability. 
     
     
         8 . The method of  claim 1 , wherein the indication regarding existence of the defect comprises a determination by an optical measuring tool or operator input or determined from yield data or electronic testing data. 
     
     
         9 . The method of  claim 1 , wherein the indication regarding existence of the defect comprises a determination by an empirical or computational model, or determination by a user of the lithographic apparatus, or a determination after patterning a layout on each die of a substrate or each substrate. 
     
     
         10 . The method of  claim 1 , wherein the classification model involves logistic regression, kernel logistic regression, support vector machine or import vector machine. 
     
     
         11 . The method of  claim 1 , wherein a number of categories of the classification model is two. 
     
     
         12 . The method of  claim 11 , wherein the categories comprise existence of defects and non-existence of defects. 
     
     
         13 . The method of  claim 1 , wherein the defects are one or more selected from a group consisting of necking, line pull back, line thinning, CD, overlapping and bridging. 
     
     
         14 . A method of training a classification model, the method comprising:
 predicting a defect in or on a substrate using the classification model, the classification model having, as an independent variable, a process parameter of a device manufacturing process for lithographically exposed substrates and/or a layout parameter of a pattern to be provided on a substrate using a lithographic apparatus;   receiving information regarding existence of a defect for a measured or determined value of the process parameter and/or layout parameter; and   training the classification model based on the predicted defect and the information regarding existence of the defect for the measured or determined value of the process parameter and/or layout parameter.   
     
     
         15 . A computer-implemented method of producing a classification model to facilitate defect prediction in a device manufacturing process involving production substrates processed by a lithographic apparatus, the method comprising training the classification model using a training set comprising measured or determined values of a process parameter of a plurality of substrates processed by the device manufacturing process and an indication regarding existence of defects associated with the values of the process parameter.