Automated model building and batch model building for a manufacturing process, process monitoring, and fault detection

• US 8,494,798 B2
• Filed: 09/02/2008
• Issued: 07/23/2013
• Est. Priority Date: 09/02/2008
• Status: Active Grant

• Current Assignee: MKS Instruments AB (Sartorius AG)
• Original Assignee: MKS Instruments Incorporated
• Inventors: Kettaneh, Nouna, Wold, Svante Bjarne
• Primary Examiner(s): BHAT, ADITYA S
• Application Number:
  US12/202,848
• Publication Number:
  US 20100057237A1
• Time in Patent Office: 1,785 Days
• Field of Search: 702/83
• US Class Current: 702/83
• CPC Class Codes: G05B 19/41885 characterised by modeling, simulation of the manufacturing system
• CPC Class Codes: G05B 2219/32075 Predict workpiece measurements from measurements of previous workpieces
• CPC Class Codes: G05B 2219/32076 Adjust feedback from previous processes as function of elapsed time
• CPC Class Codes: G05B 2219/45031 Manufacturing semiconductor wafers

Abstract

A method for creating a new model of a manufacturing process according to a multivariate analysis including selecting a set of data representative of multidimensional data measured during a step or phase of a manufacturing process. The method also includes determining a set of model generation conditions based on the set of data and generating the new model specifying intervals for the multidimensional data measured during a future manufacturing process based on the set of model generation conditions.

Claims

1. A method for creating a new model of a manufacturing process according to a multivariate analysis, the method comprising:

receiving, by a computing device, a plurality of sets of data, wherein the plurality of sets of data are multidimensional data measured during a step or phase of a manufacturing process for one or a plurality of wafers, and the step or phase is a process step from a plurality of process steps associated with manufacturing a wafer;

determining, by the computing device, a set of model generation conditions based on a user input selection that indicates which data from the plurality of sets of data to include in generating a new model, wherein the user input selection includes a maturity variable that indicates a degree of completion of a batch; and

generating, by the computing device, the new model based on the set of model generation conditions, wherein the new model is used for monitoring during a future manufacturing process.

2. The method of claim 1, further comprising:

associating a set identifier or batch ID with each set of data of the plurality of sets of data and a step identifier or phase ID with the step of the manufacturing process during which each set of data of the plurality of sets of data was measured; and

associating a time variable with each set of data of the plurality of sets of data.

3. The method of claim 1, further comprising excluding data not to be used during generation of the new model.

4. The method of claim 3, wherein excluding data comprises removing outliers from each set of data of the plurality of sets of data before the new model is generated.

5. The method of claim 3, wherein excluding the data comprises removing data that exceeds threshold parameters based on statistical calculations.

6. The method of claim 5, wherein the statistical calculations comprise at least one of a multivariate principal component analysis or partial least squares analysis.

7. The method of claim 6, wherein the multivariate principal component analysis or partial least squares analysis includes a Hotelling T²-type calculation, a DModX calculation, or a combination thereof.

8. The method of claim 1, wherein each set of data of the plurality of sets of data includes raw data from one or more tools or data from an existing model.

9. The method of claim 1, wherein selecting further comprises importing raw data from one or more process tools or importing data from a previously-generated model.

10. The method of claim 9, further comprising scaling and centering of the imported data based on a unit variance or other statistical calculation.

11. The method of claim 9, wherein the imported data are adjusted according to a user-specified set of scaling and centering values before generation of the new model.

12. The method of claim 1, wherein the model generation conditions further comprise included variables, excluded variables, included process steps, excluded process steps, time variables, or any combination thereof.

13. The method of claim 1, wherein at least one maturity variable is smoothed or linearized, or the time variables are normalized or both.

14. The method of claim 1, wherein the measured multidimensional data satisfying a threshold condition triggers generating the new model.

15. The method of claim 14, wherein the threshold condition is satisfied when a multivariate analysis of the multidimensional data results in a value that exceeds a threshold value.

16. The method of claim 15, wherein the multivariate analysis involves a Hotelling T²-type calculation, a DModX-type calculation, a weighted moving average-type calculation, or a multivariate control chart calculation, any of which are derived from a principal component analysis or a partial least squares analysis.

17. An apparatus for generating data structures indicative of conditions of a manufacturing process or current quality of output of a manufacturing process, the apparatus comprising a processor and memory configured to:

receive a plurality of sets of data, wherein the plurality of sets of data are multidimensional data measured during a step or phase of a manufacturing process for one or a plurality of wafers, and the step or phase is a process step from a plurality of process steps associated with manufacturing a wafer;

determine a set of model generation conditions based on a user input selection that indicates which data from the plurality of sets of data to include in generating a new model, wherein the user input selection includes a maturity variable that indicates a degree of completion of a batch; and

generate a new model based on the set of model generation conditions, wherein the new model is used for monitoring during a future manufacturing process.

18. A system for creating or updating a model according to a multivariate analysis, the system comprising:

a user interface including;

(a) a first window allowing a user to select a plurality of sets of data, wherein the plurality of sets of data are multidimensional data measured during a step or phase of a manufacturing process for one or a plurality of wafers, and the step or phase is a process step from a plurality of process steps associated with manufacturing a wafer, to be used to generate a new model;

(b) a second window allowing a user to determine model generation specifications including data to be included in the generated model, data to be excluded from the generated model, and a maturity variable that indicates a degree of completion of a batch; and

(c) an area for displaying the data to be used to generate the model; and

a memory for storing the generated model, the generated model based on the set of model generation specifications, wherein the generated model is used for monitoring during a future manufacturing process.

19. The system of claim 18, further comprising a third window including a user-configurable spreadsheet portion to specify scaling, centering, or pre-processing values, procedures, or transformations for applying to the data before the model is generated.