Reticle design inspection system

• US 20030086081A1
• Filed: 05/06/2002
• Published: 05/08/2003
• Est. Priority Date: 09/17/1998
• Status: Active Grant

• Current Assignee: Applied Materials Incorporated
• Original Assignee: Applied Materials Incorporated
• Inventors: Lehman, Yonatan
• Granted Patent: US 7,079,235 B2
• Time in Patent Office: Days
• US Class Current: 356/237.1
• CPC Class Codes: G01N 21/95607 using a comparative method
• CPC Class Codes: G03F 1/44 Testing or measuring features, e.g. grid patterns, focus monitors, sawtooth scales or notched scales
• CPC Class Codes: G03F 1/84 Inspecting
• CPC Class Codes: G03F 7/70433 Layout for increasing efficiency or for compensating imaging errors, e.g. layout of exposure fields for reducing focus errors; Use of mask features for increasing efficiency or for compensating imaging errors

Abstract

A method of reticle inspection, comprising generating a test reticle comprising a plurality of test pattern-features thereon; manufacturing a wafer using the reticle; and determining a transfer of at least one of said plurality of pattern features from said reticle to said wafer. Preferably, a neural network is trained using the determination. Preferably, a reticle is inspected by running detected defects through the neural network to determine if the detected defect has a consequence.

Claims

1. A method of reticle inspection, comprising:

generating a test reticle comprising a plurality of test pattern-features thereon;

manufacturing a wafer using the reticle; and

determining a transfer of at least one of said plurality of pattern features from said reticle to said wafer.

2. A method according to claim 1, comprising training a reticle inspection system to associate a test pattern feature with a consequence of said transfer.

3. A method according to claim 2, wherein training comprises training a neural network.

4. A method according to claim 1, wherein said test reticle does not define an operational layer of an operational integrated circuit.

9. A method according to claim 1, wherein determining comprises determining a consequence of said transfer.

10. A method according to claim 1, wherein said plurality of pattern-features comprise optical proximity correction patterns.

11. A method according to claim 10, wherein said plurality of test pattern features comprises a plurality of pattern-features corresponding to a single type of defect at a plurality of degrees of severity.

12. A method according to claim 10, wherein said plurality of test pattern features comprises a plurality of pattern-features corresponding to a plurality of different defect types.

13. A method according to claim 10, wherein said plurality of test pattern features comprises a plurality of copies of a single pattern-feature.

14. A method according to claim 1, comprising:

inspecting a manufactured reticle to detect one or more pattern defects; and

comparing each of said detected pattern defects to said test pattern features.

15. A method according to claim 14, comprising approving said manufactured reticle if none of said test pattern features, to which detected defects are compared and match, transfers.

5. A method according to darn 1, wherein said pattern feature represents a printing on a reticle which is supposed to meet a design rule and wherein determining a transfer comprises determining if the pattern-feature transferred in accordance with the design rule.

6. A method according to claim 5, wherein said pattern feature represents a defect in manufacturing a reticle in accordance with said design rule.

7. A method according to claim 6, wherein a defect-free pattern feature is provided on said reticle and associated with said defect representing pattern-feature.

8. A method according to claim 7, wherein determining a transfer comprises comparing an image of said defect-free feature with an image of said defect representing pattern-feature.

16. A method of reticle inspection, comprising:

providing a reticle; and

matching one or more pattern-features of said reticle to a database of defect pattern-features, each defect pattern-feature associated with a consequence of said defect feature.

17. A method according to claim 16, comprising inspecting said reticle to determine one or more possible defects in said reticle, wherein said matching comprises one or more pattern-feature which correspond to said possible defects.

18. A method according to claim 17, wherein said inspecting comprises one of:

a. comparing said reticle against a design of said reticle;

or

b. comparing features in one die of said reticle to a respective feature in an adjacent die on said reticle.

19. A method according to claim 16, wherein said consequence comprises a transfer of a defect onto a wafer manufactured using said reticle.

20. A method according to claim 16, wherein said consequence comprises a non-transfer of said defect.

21. A method according to claim 16, wherein said consequence comprises an effect on a functionality of a wafer manufactured using said reticle.

22. A method according to claim 16, wherein said matching comprises matching said one or more pattern-features to a database associated with a particular wafer manufacture process.

23. A method according to claim 16, wherein said matching comprises matching said one or more pattern-features to a database associated with a particular mask manufacture process.

24. A method according to claim 16, wherein said matching comprises matching using a neural network, which database is embodied by said neural network.

25. A reticle inspection system comprising:

a training engine receiving data of known test defects produced on a test reticle and detected defects found on a test wafer, and providing training results correlating said detected defects to said test defects to identify test defects which have been transferred to the test wafer and test defects which have not been transferred to the test wafer;

an inspection result module receiving manufacturing defects data of defects detected on said production reticle by an inspection system;

a print/no print module receiving the manufacturing defects data and the training results, and providing a defect report identifying which of the manufacturing defects is likely to be transferred to a manufactured wafer.

26. A system according to claim 25, wherein said print/no print module comprises a neural network.

27. A system according to claim 25, wherein said print/no print module comprises a pattern matching computer.

28. The reticle inspection system of claim 25, further comprising a defect review stage and a re-train feedback loop between said defect review stage and said training engine.

29. A test reticle comprising:

a reticle substrate; and

a plurality of feature-patterns formed thereon,

wherein said plurality of feature patterns comprises defect feature patterns which may transfer to a manufactured wafer.

30. A reticle according to claim 29, wherein said defects comprise defects in a reticle generation process.

31. A reticle according to claim 29, further comprising a plurality of optical proximity correction (OPC) feature-patterns.

32. A reticle according to claim 29, further comprising a plurality of defect-free pattern-features.

33. A reticle according to claim 32, wherein each of said plurality of defect-free pattern features is associated with a defect pattern-feature, which describes a same feature of a reticle.

34. A reticle according to claim 29, wherein said defects comprise defects in a wafer manufacture process.

35. A reticle according to claim 34, wherein feature patterns of said plurality of feature Patterns are selected responsive to a particular wafer manufacture process.

36. A reticle according to claim 34, wherein feature patterns of said plurality of feature patterns are selected responsive to a particular mask manufacture process.

37. A reticle according to claim 29, wherein feature patterns of said plurality of feature patterns comprise a plurality of copies of a single feature pattern.

38. A reticle according to claim 29, wherein feature patterns of defect feature patterns comprise a plurality feature patterns, each corresponding to a different defect type.

39. A reticle according to claim 29, wherein feature patterns of defect feature patterns comprise a plurality feature patterns, each corresponding to a different severity level for a single defect type.

40. A reticle inspection system, comprising:

a database associating a plurality of feature patterns with a consequence of said feature pattern;

an image input; and

a matcher which matches said input image with said database; and

a consequence output which outputs a consequence associated with said image input.

41. A system according to claim 40, wherein said database and said matcher are embodied as a neural network.

42. A system according to claim 40, wherein said consequence comprises a transfer of said feature pattern as a defect, from a reticle to a wafer.

43. A system according to claim 40, wherein said consequence is associated with a wafer manufacture process definition.

44. A system according to claim 43, wherein said process definition comprises a process window.

45. A system according to claim 40, comprising a trainer, which maintains said database.

46. A system according to claim 40, comprising a trainer, which generates said database.

47. A system according to claim 45, wherein said trainer comprises:

a defect image input;

a reference image input; and

a database interface which updates said database with a consequence, responsive to one or more differences between said defect image and said reference image.

48. A system according to claim 47, comprising a consequence input for entering a consequence.

49. A system according to claim 47, wherein said consequence comprises an indication if said defect transfers to a wafer or not.

50. A system according to claim 49, wherein said consequence is associated with one or more wafer manufacturing parameters.

51. A system according to claim 49, wherein said consequence comprises an effect on a process window for wafer manufacture using a reticle having said feature patterns thereon.

52. A system according to claim 47, wherein said image input and said reference input receive input from a single wafer scanner.

53. A system according to claim 52, wherein said scanner is a dual-headed scanner.