Method for adaptive sampling in examining an object and system thereof

• US 10,190,991 B2
• Filed: 11/03/2016
• Issued: 01/29/2019
• Est. Priority Date: 11/03/2016
• Status: Active Grant

• Current Assignee: Applied Materials Israel Limited (Applied Materials Incorporated)
• Original Assignee: Applied Materials Israel Limited (Applied Materials Incorporated)
• Inventors: Sofer, Yotam, Kaizerman, Idan
• Primary Examiner(s): Le, Toan
• Application Number:
  US15/343,090
• Publication Number:
  US 20180306728A1
• Time in Patent Office: 817 Days
• CPC Class Codes: G01N 2021/8867 using sequentially two or more inspection runs, e.g. coarse and fine, or detecting then analysing
• CPC Class Codes: G01N 2021/8877 Proximity analysis, local statistics
• CPC Class Codes: G01N 21/9501 Semiconductor wafers manufacturing processes per se of semiconductor devices implementing a measuring step H01L22/10
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: H01L 22/12 for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions electrical measurement of diffusions H01L22/14
• CPC Class Codes: H01L 22/20 Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

Examining an object, comprising: receiving potential defects, each associated with a location; performing first clustering of the potential defects to obtain first and second subsets, the clustering performed such that potential defects in the first subset are denser in a physical area than potential defects in the second subset; automatically assigning first validity probabilities to potential defects in the first and second subsets; selecting for review potential defects from the first and second subsets, according to a third policy, and in accordance with a strategy for combining top elements and randomly selected elements from the merged list; receiving indications for potential defects in part of the potential defect lists, subsequent to potential defects being reviewed; updating the policies in accordance with validation or classification of items in the first and second subsets; and repeating said assigning, selecting, receiving and updating with the updated policies, until a stopping criteria is observed.

Claims

1. A method of examining an object, the method comprising:

(a) identifying, by a processor, a plurality of potential defects, each potential defect of the plurality of potential defects being associated with a potential defect location;

(b) performing, by the processor, a first clustering of the plurality of potential defects to obtain a first subset and one or more second subsets, the clustering performed in accordance with spatial distances between potential defect locations such that potential defects in the first subset are characterized by higher density in at least one physical area than potential defects in the one or more second subsets;

(c) assigning, by the processor, first probabilities to potential defects in the first subset to be valid defects, the first probabilities being calculated in accordance with a first policy;

(d) assigning, by the processor, second probabilities to potential defects in the one or more second subsets to be valid defects, the second probabilities being calculated in accordance with one or more second policies specifying how to combine at least two second factors associated with potential defect locations;

(e) selecting, by the processor, at least one potential defect from the first subset and the one or more second subsets for review by a review tool in accordance with a third policy specifying how to combine potential defects from a plurality of subsets into a merged list, and wherein the selecting of the at least one potential defect for review by the review tool is further in accordance with a strategy indicating how to combine top elements from the merged list and randomly selected elements from the merged list in accordance with a number of reviews from the review tool;

(f) receiving, by the processor, validity or class indications for potential defects in a potential defect lists associated with the selected at least one potential defect, the validity or class indications being received subsequent to potential defects in the potential defect lists being reviewed by the review tool; and

(g) subsequent to a stopping criteria not being observed;

(i) updating, by the processor, the first, second, or third policy in accordance with validation or classification of an item in the first subset and the one or more second subsets; and

(ii) repeating, by the processor, steps (c)-(g) in accordance with the first, second, or third policy as updated, until the stopping criteria is observed.

2. The method of claim 1, wherein the stopping criteria is selected from at least one of:

a predetermined number of potential defects being reviewed by the review tool, convergence of a number of indications that a potential defect is valid and a predetermined number of defects associated with a specific defect class being detected.

3. The method of claim 1, wherein a decision confidence level is received for at least one of the validity or class indications.

4. The method of claim 1, wherein the at least one physical area is selected from at least one of:

a region, an area in the vicinity of an open line, and an area in the vicinity of a closed line.

5. The method of claim 1, further comprising:

performing a second clustering of the one or more second subsets set based on characteristics other than geometric location.

6. The method of claim 5, wherein the characteristics of the second clustering include at least one item selected from:

shade or color of a wafer at a location of a potential defect, background shade or color of the wafer at an area of the potential defect, edges in the vicinity of the potential defect, a feature in the vicinity of the potential defect, number of polygons, edges, or corners and density thereof in the vicinity of the potential defect.

7. The method of claim 5, wherein the second clustering provides for determining potential defects located in a dense area in relation to characteristics used for the second clustering, thereby providing for determining systematic potential defects.

8. The method of claim 5, wherein the at least two second factors are selected from at least one of:

being located at an area which is sparse in feature space in relation to the characteristics used for the second clustering;

being located at an outlier area of an area which is dense in feature space in relation to the characteristics used for the second clustering;

proximity to a validated potential defect; and

a combination of two or more of the above,

thereby providing for determining random potential defects.

9. The method of claim 8, wherein the random potential defects are associated with at least two scans taken by different tools or under different conditions.

10. The method of claim 1, wherein

selecting for review comprises:

sorting potential defects of the first subset according to at least one first probability to obtain a first sorted list;

sorting the one or more second subsets according to at least one second probability to obtain at least one second sorted list;

prioritizing potential defects from the first sorted list or the second sorted list to form a merged list in accordance with the third policy; and

selecting potential defects to be reviewed from the merged list in accordance with the strategy.

11. The method of claim 10, wherein the third policy complies with at least one item selected from at least one of:

selecting a predetermined percentage of a required number of potential defects from the first sorted list and complementing with potential defects from the second sorted list;

selecting top probability potential defects from a merged list of the first sorted list and the second sorted list, sorted by probability; and

prioritizing potential defects from the first sorted list or the second sorted list in accordance with user input and selecting top priority potential defects.

12. The method of claim 10, wherein the strategy indicates a number of top probability potential defects to be selected from the merged list, and complementing with randomly selected potential defects from the merged list.

13. The method of claim 12, wherein the strategy changes between iterations.

14. The method of claim 12, wherein the strategy changes between iterations in accordance with an item selected from at least one of:

increasing the number of randomly selected potential defects in advanced iterations relative to earlier iterations, increasing the number of randomly selected potential defects if a number of true defects identified in a previous iteration is below a threshold, increasing the number of randomly selected potential defects if a number of true defects identified in a previous iterations converges, and changing the number of randomly selected potential defects in accordance with type or location distribution of true defects identified in one or more previous iterations.

15. The method of claim 1, wherein assigning the first probabilities or assigning the second probabilities comprises receiving a user indication to assigning higher priority to potential defects at specific areas or of specific classes.

16. The method of claim 1, wherein the object is a wafer or a mask.

17. A system comprising:

a memory; and

a processor, operatively coupled with the memory, to;

(a) identify a plurality of potential defects, each potential defect of the plurality of potential defects being associated with a potential defect location;

(b) perform a first clustering of the plurality of potential defects to obtain a first subset and one or more second subsets, the clustering performed in accordance with spatial distances between potential defect locations such that potential defects in the first subset are characterized by higher density in at least one physical area than potential defects in the one or more second subsets;

(c) assign first probabilities to potential defects in the first subset to be valid defects, the first probabilities being calculated in accordance with a first policy;

(d) assign second probabilities to potential defects in the one or more second subsets to be valid defects, the second probabilities being calculated in accordance with one or more second policies specifying how to combine at least two second factors associated with potential defect locations;

(e) select at least one potential defect from the first subset and the one or more second subsets for review by a review tool in accordance with a third policy specifying how to combine potential defects from a plurality of subsets into a merged list, and wherein the selecting of the at least one potential defect for review by the review tool is further in accordance with a strategy indicating how to combine top elements from the merged list and randomly selected elements from the merged list in accordance with a number of reviews from the review tool;

(f) receive validity or class indications for potential defects in a potential defect lists associated with the selected at least one potential defect, the validity or class indications being received subsequent to potential defects in the potential defect lists being reviewed by the review tool; and

(g) subsequent to a stopping criteria not being observed;

(i) update the first, second, or third policy in accordance with validation or classification of an item in the first subset and the one or more second subsets; and

(ii) repeat steps (c)-(g) in accordance with the first, second, or third policy as updated, until the stopping criteria is observed.

18. The system of claim 17, wherein the stopping criteria is selected from one or more of:

a predetermined number of potential defects being reviewed by the review tool, convergence of a number of indications that a potential defect is valid, and a predetermined number of defects associated with a specific defect class being detected.

19. The computerized system of claim 17, wherein to select the at least one potential defect, the processor is further to:

sort potential defects of the first subset according to at least one first probability to obtain a first sorted list;

sort the one or more second subsets according to at least one second probability to obtain a second sorted list;

prioritize potential defects from the first sorted list or the second sorted list to form a merged list in accordance with the third policy; and

select potential defects to be reviewed from the merged list in accordance with the strategy, and wherein the third policy complies with at least one item selected from;

selecting a predetermined percentage of a required number of potential defects from the first sorted list and complementing with potential defects from the second sorted list;

selecting top probability potential defects from a merged list of the first sorted list and the second sorted list, sorted by probability; and

prioritizing potential defects from the first sorted list or the second sorted list in accordance with user input and selecting top priority potential defects.

20. A non-transitory computer readable storage medium retaining program instructions, which when read by a processor, cause the processor to perform operations comprising:

(a) identifying, by the processor, a plurality of potential defects, each potential defect of the plurality of potential defects being associated with a potential defect location;

(b) performing, by the processor, a first clustering of the plurality of potential defects to obtain a first subset and one or more second subsets, the clustering performed in accordance with spatial distances between potential defect locations such that potential defects in the first subset are characterized by higher density in at least one physical area than potential defects in the one or more second subsets;

(c) assigning, by the processor, first probabilities to potential defects in the first subset to be valid defects, the first probabilities being calculated in accordance with a first policy;

(d) assigning, by the processor, second probabilities to potential defects in the one or more second subsets to be valid defects, the second probabilities being calculated in accordance with one or more second policies specifying how to combine at least two second factors associated with potential defect locations;

(e) selecting, by the processor, at least one potential defect from the first subset and the one or more second subsets for review by a review tool in accordance with a third policy specifying how to combine potential defects from a plurality of subsets into a merged list, and wherein the selecting of the at least one potential defect for review by the review tool is further in accordance with a strategy indicating how to combine top elements from the merged list and randomly selected elements from the merged list in accordance with a number of reviews from the review tool;

(f) receiving, by the processor, validity or class indications for potential defects in a potential defect lists associated with the selected at least one potential defect, the validity or class indications being received subsequent to potential defects in the potential defect lists being reviewed by the review tool; and

(g) subsequent to a stopping criteria not being observed;

(i) updating, by the processor, the first, second, or third policy in accordance with validation or classification of an item in the first subset and the one or more second subsets; and

(ii) repeating, by the processor, steps (c)-(g) in accordance with the first, second, or third policy as updated, until the stopping criteria is observed.