Multi-variable regression for metrology
View Patent ↗A method for assessing metrology tool accuracy is described. Multi-variable regression is used to define the accuracy of a metrology tool such that the interaction between different measurement parameters is taken into account. A metrology tool under test (MTUT) and a reference metrology tool (RMT) are used to measure a set of test profiles. The MTUT measures the test profiles to generate a MTUT data set for a first measurement parameter. The RMT measures the test profiles to generate RMT data sets for the first measurement parameter, and at least a second measurement parameter. Multi-variable regression is then performed to generate a best-fit plane for the data sets. The coefficient of determination (R 2 value) represents the accuracy index of the MTUT.
1. A method for assessing a metrology tool under test (MTUT) comprising:
measuring a target parameter from a plurality of test structures using the MTUT to generate a test data set, the test data set corresponds to a test variable;
measuring a first reference parameter from the plurality of test structures using a reference metrology tool (RMT) to generate a first reference data set, the first reference data set corresponds to a first reference variable;
measuring a second reference parameter, which is different from the first reference parameter, from the plurality of test structures using the RMT to generate a second reference data set, the second reference data set corresponds to a second reference variable, wherein the first and second reference parameters measured by the RMT are compared with the target parameter measured by the MTUT, and wherein the first reference parameter is the same as the target parameter and the second reference parameter influences the target parameter; and
performing a multi-variable regression analysis, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and first and second reference variables of the first and second reference data sets from the RMT.
2. The method of claim 1 wherein the MTUT comprises a first type metrology tool and the RMT comprises a second type metrology tool.
3. The method of claim 2 wherein the first and second types are selected from AFM, SCM, SEM and TEM.
4. The method of claim 2 wherein the first and second types are different types.
5. The method of claim 4 wherein the first and second types are selected from AFM, SCM, SEM and TEM.
6. The method of claim 4 wherein the multi-variable regression analysis comprises ordinary least squares or Mandel.
7. The method of claim 4 wherein the multi-variable regression analysis determines a coefficient of determination, wherein the coefficient of determination represents the accuracy index of the MTUT.
8. The method of claim 1 wherein the multi-variable regression analysis determines a coefficient of determination, wherein the coefficient of determination represents the accuracy index of the MTUT.
9. The method of claim 1 wherein the multi-variable regression analysis comprises ordinary least squares or Mandel.
10. The method of claim 9 wherein the multi-variable regression analysis determines a coefficient of determination, wherein the coefficient of determination represents the accuracy index of the MTUT.
11. The method of claim 1 wherein the target and the first and second reference parameters are selected from critical dimension, sidewall angles, top corner rounding, profile height, underneath film thickness or a combination thereof.
12. The method of claim 1 wherein the target parameter comprises critical dimension of the test structures and the first and second reference parameters comprise critical dimension and sidewall angle of the test structures.
13. The method of claim 1 comprises:
measuring more than two different reference parameters with the RMT to generate more than two different reference data sets corresponding to two or more different reference variables; and
performing a multi-variable regression analysis, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and more than two different reference variables of the more than two different reference data sets from the RMT.
14. A method of fabricating a device comprising:
assessing a metrology tool under test (MTUT) by
measuring a target parameter from a plurality of test structures using the MTUT to generate a test data set, the test data set corresponds to a test variable;
measuring a first reference parameter from the plurality of test structures using a reference metrology tool (RMT) to generate a first second reference data set,
measuring a second reference parameter from the plurality of test structures using the RMT to generate a second reference data set, wherein the first and second reference parameters measured by the RMT are compared with the target parameter measured by the MTUT, and wherein the first reference parameter is the same as the target parameter and the second reference parameter influences the target parameter; and
performing a multi-variable regression analysis, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and first and second reference variables of the first and second reference data sets from the RMT;
using the assessed MTUT to characterize a feature during processing;
determining processing parameters for a processing tool based on the characterized feature; and
processing a wafer on the processing tool based on the processing parameters to form the device.
15. The method of claim 14 wherein the MTUT comprises a first type metrology tool and the RMT comprises a second type metrology tool selected from AFM, SCM, SEM and TEM.
16. The method of claim 14 wherein the multi-variable regression analysis determines a coefficient of determination, wherein the coefficient of determination represents the accuracy index of the MTUT.
17. The method of claim 14 wherein the multi-variable regression analysis comprises ordinary least squares or Mandel.
18. The method of claim 14 wherein the target and the first and second reference parameters are selected from critical dimension, sidewall angles, top corner rounding, profile height, underneath film thickness or a combination thereof.
19. The method of claim 14 wherein the target parameter comprises critical dimension of the test structures and the first and second reference parameters comprise critical dimension and sidewall angle of the test structures.
20. The method of claim 14 comprises:
measuring more than two different reference parameters with the RMT to generate more than two different reference data sets corresponding to two or more different reference variables; and
performing a multi-variable regression analysis, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and more than two different reference variables of the more than two different reference data sets from the RMT.
21. A method for increasing accuracy of assessing a metrology tool under test (MTUT) comprising:
measuring a target parameter from a plurality of test structures using the MTUT to generate a test data set, the test data set corresponds to a test variable;
measuring a first reference parameter from the plurality of test structures using a reference metrology tool (RMT) to generate a first reference data set;
measuring a second reference parameter from the plurality of test structures using the RMT to generate a second reference data set, wherein the first and second reference parameters measured by the RMT are compared with the target parameter measured by the MTUT, and wherein the first reference parameter is the same as the target parameter and the second reference parameter influences the target parameter; and
performing a multi-variable regression analysis to determine a coefficient of determination which represents the accuracy index of the MTUT, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and first and second reference variables of the first and second reference data sets from the RMT.
22. The method of claim 21 wherein the MTUT comprises a first type metrology tool and the RMT comprises a second type metrology tool selected from AFM, SCM, SEM and TEM.
23. The method of claim 21 comprises:
measuring more than two different reference parameters with the RMT to generate more than two different reference data sets corresponding to two or more different reference variables; and
performing a multi-variable regression analysis, wherein a multi-variable model of the multi-variable regression analysis is based on the test variable of the test data set from the MTUT and more than two different reference variables of the more than two different reference data sets from the RMT.