Predictive wafer modeling based focus error prediction using correlations of wafers

Patent metadata

Field	Value
Publication number	US-9707660-B2
Application number	US-201414457706-A
Country	US
Kind code	B2
Filing date	Aug 12, 2014
Priority date	Apr 22, 2014
Publication date	Jul 18, 2017
Grant date	Jul 18, 2017

Abstract

Predictive modeling based focus error prediction method and system are disclosed. The method includes obtaining wafer geometry measurements of a plurality of training wafers and grouping the plurality of training wafers to provide at least one training group based on relative homogeneity of wafer geometry measurements among the plurality of training wafers. For each particular training group of the at least one training group, a predictive model is develop utilizing non-linear predictive modeling. The predictive model establishes correlations between wafer geometry parameters and focus error measurements obtained for each wafer within that particular training group, and the predictive model can be utilized to provide focus error prediction for an incoming wafer belonging to that particular training group.

First claim

What is claimed is:

1. A predictive modeling based focus error prediction method, comprising:
obtaining wafer geometry measurements of a plurality of training wafers;
grouping the plurality of training wafers to provide at least one training group based on relative homogeneity of wafer geometry measurements among the plurality of training wafers;
for each particular training group of the at least one training group, developing a predictive model for that particular training group utilizing non-linear predictive modeling, the predictive model establishing correlations between wafer geometry parameters and focus error measurements obtained for each wafer within that particular training group; and
utilizing the predictive model developed for a particular training group of the at least one training group to provide focus error prediction for an incoming wafer belonging to that particular training group.

2. The method of claim 1 , wherein the grouping step further comprises:
calculating correlations between each pair of wafers within the plurality of training wafers to generate a pair-wise correlation matrix;
converting the pair-wise correlation matrix to a binary relationship matrix based on whether each pair-wise correlation is greater than a predetermined similarity threshold;
identifying a maximal sub matrix containing all binary values of one from the binary relationship matrix;
identifying the wafers corresponding to the identified maximal sub matrix as being in a homogenous group;
removing the identified maximal sub matrix from the binary relationship matrix; and
repeating the steps of: identifying a maximal sub matrix containing all binary values of one from the binary relationship matrix and identifying the wafers corresponding to the identified maximal sub matrix as being in a homogenous group.

3. The method of claim 2 , wherein the correlations between each pair of wafers within the plurality of training wafers are calculated at least partially based on wafer flatness measurements of the plurality of training wafers.

4. The method of claim 1 , further comprising:
validating the predictive model developed for a particular training group of the at least one training group prior to utilizing the predictive model for focus error prediction.

5. The method of claim 4 , wherein validating the predictive model developed for a particular training group further comprises:
validating the predictive model based on at least one wafer within the particular training group.

6. The method of claim 4 , wherein validating the predictive model developed for a particular training group further comprises:
validating the predictive model based on at least one wafer outside of the particular training group.

7. The method of claim 1 , further comprising:
providing the focus error prediction for the incoming wafer to control a wafer process tool.

8. The method of claim 7 , wherein the wafer process tool is a lithography process tool.

9. The method of claim 1 , wherein the non-linear predictive modeling utilizes at least one of: a neural network, a random forest, a boosted regression tree and a support vector machine.

10. The method of claim 1 , wherein said grouping the plurality of training wafers is based on relative homogeneity of wafer geometry measurements among the plurality of training wafers and further based on relative homogeneity of overlay errors among the plurality of training wafers.

11. A method, comprising:
obtaining wafer geometry measurements of a plurality of wafers;
calculating correlations between each pair of wafers within the plurality of wafers to generate a pair-wise correlation matrix based on wafer geometry measurements;
converting the pair-wise correlation matrix to a binary relationship matrix based on whether each pair-wise correlation is greater than a predetermined similarity threshold;
dividing the plurality of wafers into a plurality of wafer groups based on relative homogeneity among the plurality of wafers, further comprising:
identifying a maximal sub matrix containing all binary values of one from the binary relationship matrix;
identifying the wafers corresponding to the identified maximal sub matrix as being in a homogenous group;
removing the identified maximal sub matrix from the binary relationship matrix; and
repeating the steps of: identifying a maximal sub matrix containing all binary values of one from the binary relationship matrix and identifying the wafers corresponding to the identified maximal sub matrix as being in a homogenous group.

12. The method of claim 11 , wherein the correlations between each pair of wafers within the plurality of training wafers are calculated at least partially based on wafer flatness measurements of the plurality of training wafers.

13. The method of claim 11 , further comprising:
for each particular wafer group of the plurality of wafer groups, developing a predictive model for that particular wafer group utilizing non-linear predictive modeling, the predictive model establishing correlations between wafer geometry parameters and focus error measurements obtained for each wafer within that particular wafer group.

14. The method of claim 13 , further comprising:
validating the predictive model developed for a particular wafer group.

15. The method of claim 14 , wherein validating the predictive model developed for a particular wafer group further comprises:
validating the predictive model based on at least one wafer within the particular wafer group.

16. The method of claim 14 , wherein validating the predictive model developed for a particular wafer group further comprises:
validating the predictive model based on at least one wafer outside of the particular wafer group.

17. The method of claim 13 , further comprising:
utilizing the predictive model developed for a particular wafer group to provide focus error prediction for an incoming wafer belonging to that particular training group.

18. The method of claim 17 , further comprising:
providing the focus error prediction for the incoming wafer to control a wafer process tool.

19. The method of claim 18 , wherein the wafer process tool is a lithography process tool.

20. A system, comprising:
a wafer geometry measurement tool configured to obtain wafer geometry measurements of a plurality of training wafers; and
a processor in communication with the wafer geometry measurement tool, the processor configured to:
group the plurality of training wafers to provide at least one training group based on relative homogeneity of wafer geometry measurements among the plurality of training wafers;
for each particular training group of the at least one training group, develop a predictive model for that particular training group utilizing non-linear predictive modeling, wherein the predictive model establishes correlations between wafer geometry parameters and focus error measurements obtained for each wafer within that particular training group; and
utilize the predictive model developed for a particular training group of the at least one training group to provide focus error prediction for an incoming wafer belonging to that particular training group.

21. The system of claim 20 , wherein to group the plurality of training wafers, the processor is further configured to:
a) calculate correlations between each pair of wafers within the plurality of training wafers to generate a pair-wise correlation matrix;
b) convert the pair-wise correlation ma

Assignees

• Kla Tencor Corp

Inventors

• Vukkadala Pradeep
• Sinha Jaydeep
• Chang Wei
• Rao Krishna

Classifications

• B24B37/005 Primary

  Control means for lapping machines or devices · CPC title

• G03F9/7026

  Focusing · CPC title

• G01B21/00

  Measuring arrangements or details thereof, where the measuring technique is not covered by the other groups of this subclass, unspecified or not relevant · CPC title

• G06N20/00

  Machine learning · CPC title

• B24B49/00

  Measuring or gauging equipment for controlling the feed movement of the grinding tool or work; Arrangements of indicating or measuring equipment, e.g. for indicating the start of the grinding operation (B24B33/06, B24B37/005 take precedence; if applicable to other machine tools, B23Q15/00 - B23Q17/00 take precedence) · CPC title