STATISTICAL MODEL-BASED METROLOGY

• US 20140297211A1
• Filed: 03/24/2014
• Published: 10/02/2014
• Est. Priority Date: 03/27/2013
• Status: Active Grant

• Current Assignee: KLA-Tencor Corporation
• Original Assignee: KLA-Tencor Corporation
• Inventors: Pandev, Stilian Ivanov, Madsen, Jonathan M.
• Granted Patent: US 10,101,670 B2
• Time in Patent Office: Days
• US Class Current: 702/81
• CPC Class Codes: G03F 7/70558 Dose control, i.e. achievement of a desired dose
• CPC Class Codes: G03F 7/70641 Focus
• 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: H01L 2924/00 Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
• CPC Class Codes: H01L 2924/0002 Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00

Abstract

Methods and systems for creating a measurement model based on measured training data are presented. The trained measurement model is used to calculate process parameter values, structure parameter values, or both, directly from measured data collected from other wafers. The measurement models receive measurement data directly as input and provide process parameter values, structure parameter values, or both, as output. The measurement model enables the direct measurement of process parameters. Measurement data from multiple targets is collected for model building, training, and measurement. In some examples, the use of measurement data associated with multiple targets eliminates, or significantly reduces, the effect of under layers in the measurement result, and enables more accurate measurements. Measurement data collected for model building, training, and measurement, may be derived from measurements performed by a combination of multiple, different measurement techniques.

Claims

1. A method comprising:

receiving a first amount of measurement data associated with measurements of a first plurality of sites on a surface of a semiconductor wafer with known variations of at least one process parameter, structure parameter, or both, wherein the first amount of measurement data is derived from measurements performed by at least one metrology technique;

determining an expected response model of each of the at least one known process parameters, structure parameters, or both;

determining an input-output measurement model based at least in part on the first amount of measurement data; and

training the input-output measurement model based on parameter values determined from the expected response model.

2. The method of claim 1, further comprising:

receiving a second amount of measurement data associated with measurements of a second plurality of sites using the first metrology technique;

determining at least one process parameter value, at least one structural parameter value, or both, associated with each of the second plurality of sites based on a fitting of the second amount of measurement data to the trained input-output measurement model; and

storing any of the at least one process parameter value, the at least one structural parameter value, or both, in a memory.

3. The method of claim 1, wherein the first amount of measurement data is associated with measurements of a first plurality of sites with known variations of focus and exposure dosage over the surface of the semiconductor wafer.

4. The method of claim 1, further comprising:

extracting one or more features of the first amount of measurement data by reducing a dimension of the first amount of measurement data, and wherein the determining the input-output measurement model is based at least in part on the one or more features.

5. The method of claim 4, wherein the reducing the dimension of the first amount of measurement data involves any of a principal components analysis, a non-linear principal components analysis, a selection of individual signals from the first amount of measurement data, and a filtering of the first amount of measurement data.

6. The method of claim 1, wherein the expected response model is a wafer map model, and wherein the determining the wafer map model involves fitting a two dimensional map function to the known process parameters, structure parameters, or both, associated with the first plurality of sites.

7. The method of claim 1, further comprising:

determining the at least one known structure parameter based at least in part on a simulation of a process performed on the semiconductor wafer.

8. The method of claim 1, further comprising:

determining the at least one known structure parameter based at least in part on an amount of reference measurement data associated with the at least one known structure parameter.

9. The method of claim 1, wherein the first amount of measurement data includes measurement signals associated with more than one target feature at any of the first plurality of sites.

10. The method of claim 1, wherein the first amount of measurement data includes measurement signals associated with more than one metrology technique.

11. A system comprising:

a metrology tool including an illumination source and a detector configured to perform measurements of a target structure; and

a computing system configured to;

receive a first amount of measurement data from the metrology tool, the first amount of data associated with measurements of a first plurality of sites on a surface of a semiconductor wafer with known variations of at least one process parameter, structure parameter, or both;

determine an expected response model of each of the at least one known process parameters, structure parameters, or both;

determine an input-output measurement model based at least in part on the first amount of measurement data; and

train the input-output measurement model based on parameter values determined from the expected response model.

12. The system of claim 11, wherein the computing system is further configured to:

receive a second amount of measurement data associated with measurements of a second plurality of sites;

determine at least one process parameter value, at least one structural parameter value, or both, associated with each of the second plurality of sites based on a fitting of the second amount of measurement data to the trained input-output measurement model; and

store any of the at least one process parameter value, the at least one structural parameter value, or both, in a memory.

13. The system of claim 11, wherein the computing system is further configured to:

extract one or more features of the first amount of measurement data by reducing a dimension of the first amount of measurement data, and wherein the determining the input-output measurement model is based at least in part on the one or more features.

14. The system of claim 13, wherein the reducing the dimension of the first amount of measurement data involves any of a principal components analysis, a non-linear principal components analysis, a selection of individual signals from the first amount of measurement data, and a filtering of the first amount of measurement data.

15. The system of claim 11, wherein the expected response model is a wafer map model, and wherein the determining the wafer map model involves fitting a two dimensional map function to the known process parameters, structure parameters, or both, associated with the first plurality of sites.

16. The system of claim 11, wherein the first amount of measurement data includes measurement signals associated with more than one target feature at any of the first plurality of sites.

17. The system of claim 11, wherein the first amount of measurement data includes measurement signals associated with more than one metrology technique.

18. A method comprising:

receiving a first amount of measurement data associated with measurements of a semiconductor target on a first semiconductor wafer;

determining at least one process parameter value, at least one structural parameter value, or both, based on a fitting of the first amount of measurement data to a trained input-output measurement model; and

storing any of the at least one process parameter value, the at least one structural parameter value, or both, in a memory.

19. The method of claim 18, further comprising:

generating the trained input-output measurement model, wherein the generating involves;

receiving a second amount of measurement data associated with measurements of the semiconductor target on a second semiconductor wafer with known variations of at least one process parameter, structure parameter, or both;

determining an expected response model of each of the at least one known process parameters, structure parameters, or both;

determining an input-output measurement model based at least in part on the second amount of measurement data; and

training the input-output measurement model based on parameter values determined from the expected response model.

20. The method of claim 19, further comprising:

extracting one or more features of the second amount of measurement data by reducing a dimension of the second amount of measurement data, and wherein the determining the input-output measurement model is based at least in part on the one or more features; and

extracting one or more features of the first amount of measurement data by reducing a dimension of the first amount of measurement data, and wherein the determining the at least one process parameter value, at least one structural parameter value, or both, is based at least in part on the one or more features.