System and Method for Inspection Using Tensor Decomposition and Singular Value Decomposition

Patent metadata

Field	Value
Publication number	US-2020244963-A1
Application number	US-202016744301-A
Country	US
Kind code	A1
Filing date	Jan 16, 2020
Priority date	Jan 28, 2019
Publication date	Jul 30, 2020
Grant date	—

Abstract

A sample characterization system is disclosed. In embodiments, the sample characterization system includes a controller communicatively coupled to an inspection sub-system, the controller including one or more processors configured to execute a set of program instructions stored in memory, the set of program instructions configured to cause the one or more processors to: acquire one or more target image frames of a sample; generate a target tensor with the one or more acquired target image frames; perform a first set of one or more decomposition processes on the target tensor to form generate one or more reference tensors including one or more reference image frames; identify one or more differences between the one or more target image frames and the one or more reference image frames; and determine one or more characteristics of the sample based on the one or more identified differences.

First claim

What is claimed:

1 . A sample characterization system, comprising:
a controller communicatively coupled to an inspection sub-system, the controller including one or more processors configured to execute a set of program instructions stored in memory, the set of program instructions configured to cause the one or more processors to:
acquire one or more target image frames of a sample;
generate a target tensor with the one or more acquired target image frames;
perform a first set of one or more decomposition processes on the target tensor to form generate one or more reference tensors including one or more reference image frames;
identify one or more differences between the one or more target image frames and the one or more reference image frames; and
determine one or more characteristics of the sample based on the one or more identified differences.

2 . The system of claim 1 , wherein the first set of one or more decomposition processes comprise a Tucker decomposition process.

3 . The system of claim 2 , wherein the Tucker decomposition process comprises a multilinear singular value decomposition process.

4 . The system of claim 1 , wherein the controller is configured to perform the first set of one or more decomposition processes on the target tensor using a first orthonormal basis vector for a column space of the target tensor, a second orthonormal basis vector for a row space of the target tensor, and a third orthonormal basis vector for a stack space of the target tensor.

5 . The system of claim 1 , wherein the controller is configured to generate the one or more reference tensors by:
performing one or more low-rank approximations of a core tensor used to carry out the one or more decomposition processes.

6 . The system of claim 5 , wherein the controller is configured to perform the one or more low-rank approximations by truncating at least a portion of the core tensor.

7 . The system of claim 5 , wherein the controller is configured to perform the one or more low-rank approximations by truncating one or more orthonormal basis vectors used for the one or more decomposition processes.

8 . The system of claim 1 , wherein the one or more identified characteristics comprise a defect of the sample.

9 . The system of claim 1 , wherein the controller is further configured to:
generate one or more control signals configured to selectively adjust one or more characteristics of one or more process tools based on the one or more determined characteristics.

10 . The system of claim 1 , wherein the controller is further configured to:
perform a second set of one or more decomposition processes on the target tensor to form a de-noised core tensor; and perform one or more high-rank approximations on the de-noised core tensor to generate a de-noised target tensor, the de-noised target tensor including one or more de-noised target image frames.

11 . The system of claim 10 , wherein the controller is further configured to:
identify one or more differences between the one or more de-noised target image frames and the one or more reference image frames; and determine one or more characteristics of the sample based on the one or more identified differences.

12 . A method for characterizing a sample, comprising:
acquiring one or more target image frames of a sample; generating a target tensor with the one or more acquired target image frames; performing a first set of one or more decomposition processes on the target tensor to generate one or more reference tensors including one or more reference image frames; identifying one or more differences between the one or more target image frames and the one or more reference image frames; and determining one or more characteristics of the sample based on the one or more identified differences.

13 . The method of claim 12 , wherein the first set of one or more decomposition processes comprise a Tucker decomposition process.

14 . The method of claim 13 , wherein the Tucker decomposition process comprises a multilinear singular value decomposition process.

15 . The method of claim 12 , wherein performing the first set of one or more decomposition processes on the target tensor comprises:
performing the first set of one or more decomposition processes on the target tensor using a first orthonormal basis vector for a row space of the target tensor, a second orthonormal basis vector for a column space of the target tensor, and a third orthonormal basis vector for a stack space of the target tensor.

16 . The method of claim 12 , wherein generating one or more reference tensors including one or more reference image frames based on a core tensor comprises:
performing one or more low-rank approximations of the core tensor used to carry out the one or more decomposition processes.

17 . The method of claim 16 , wherein performing one or more low-rank approximations of the core tensor comprises:
truncating at least a portion of the core tensor.

18 . The method of claim 16 , wherein performing one or more low-rank approximations of the core tensor comprises:
truncating one or more orthonormal basis vectors used for the one or more decomposition processes.

19 . The method of claim 12 , wherein the one or more identified characteristics comprise a defect of the sample.

20 . The method of claim 12 , further comprising:
generating one or more control signals configured to selectively adjust one or more characteristics of one or more process tools based on the one or more determined characteristics.

21 . The method of claim 12 , further comprising:
performing a second set of one or more decomposition processes on the target tensor to form a de-noised core tensor; and performing one or more high-rank approximations on the de-noised core tensor to generate a de-noised target tensor, the de-noised target tensor including one or more de-noised target image frames.

22 . The method of claim 21 , further comprising:
identifying one or more differences between the one or more de-noised target image frames and the one or more reference image frames; and determining one or more characteristics of the sample based on the one or more identified differences.

23 . A sample characterization system, comprising:
a controller communicatively coupled to an inspection sub-system, the controller including one or more processors configured to execute a set of program instructions stored in memory, the set of program instructions configured to cause the one or more processors to:
acquire one or more difference image frames of a sample, the one or more difference image frames based on one or more target image frames and one or more reference image frames;
generate one or more stacked difference images with the one or more acquired difference image frames;
perform a set of one or more singular value decomposition (SVD) processes on the one or more stacked difference images to form a set of one or more singular vectors;
selectively modify at least one singular vector of the set of one or more singular vectors to generate a modified set of one or more singular vectors;
generate a modified stacked difference image based on the modified set of one or more singular vectors; and
determine one or more characteristics of the sample based on the modified stacked difference image.

24 . The system of claim 23 , wherein the controller is configured to determine o

Assignees

• Kla Corp

Inventors

• Patwary Nurmohammed
• Wallingford Richard
• Smith James A
• Li Xiaochun
• Tumakov Vladimir
• Brauer Bjorn

Classifications

• G06T2207/30148

  Semiconductor; IC; Wafer · CPC title

• G06T7/001 Primary

  using an image reference approach · CPC title

• H04N19/60

  using transform coding · CPC title

• H04N19/126 Primary

  Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers · CPC title

• G01N21/9501

  Semiconductor wafers (manufacturing processes per se of semiconductor devices implementing a measuring step H10P74/20) · CPC title