Process control techniques for semiconductor manufacturing processes

Patent metadata

Field	Value
Publication number	US-10734293-B2
Application number	US-201515604240-A
Country	US
Kind code	B2
Filing date	Nov 25, 2015
Priority date	Nov 25, 2014
Publication date	Aug 4, 2020
Grant date	Aug 4, 2020

Abstract

Techniques for measuring and/or compensating for process variations in a semiconductor manufacturing processes. Machine learning algorithms are used on extensive sets of input data, including upstream data, to organize and pre-process the input data, and to correlate the input data to specific features of interest. The correlations can then be used to make process adjustments. The techniques may be applied to any feature or step of the semiconductor manufacturing process, such as overlay, critical dimension, and yield prediction.

First claim

The invention claimed is:

1. A method, comprising:
receiving a plurality of real-time inputs as input data into an overlay measurement model stored in a data processing apparatus, a first set of the real-time inputs collected from metrology measurements in a current production run of semiconductor wafers in a lithography process and a second set of the real-time inputs collected from metrology measurements in at least one upstream process;
analyzing the input data using the overlay measurement model to determine a multi-variate relationship of a first plurality of the real-time inputs to a targeted overlay measurement;
evaluating the multi-variate relationship of the first plurality of the real-time inputs to form a prediction of the targeted overlay measurement in the lithography process for the current production run; and
adjusting the lithography process or the upstream process such that the prediction of the targeted overlay measurement correlates with an actual targeted overlay measurement.

2. The method of claim 1 , further comprising:
the overlay measurement model obtains additional input data from processes in the previous production runs after the lithography process for use in determining the multi-variate relationship; and
feeding additional real-time inputs from processes after the lithography process into the model for each production run.

3. A method, comprising:
obtaining a plurality of actual overlay measurements from a plurality of wafers in a plurality of production runs of a lithography process, wherein each actual overlay measurement indicates an offset between a first set of features formed on a first layer and a second set of features formed on a second layer above the first layer;
collecting a plurality of sets of input data, each set of input data collected from one of the plurality of production runs including data obtained from the lithography process and data obtained from at least one upstream process;
analyzing the sets of input data collected from each production run to determine a multi-variate relationship of the input data to the actual overlay measurements;
evaluating the multi-variate relationship to generate a predicted overlay measurement for each set of input data; and
adjusting the lithography process or the at least one upstream process such that the predicted overlay measurements correlate with the actual overlay measurements for each set of input data.

4. The method of claim 3 , further comprising:
creating a model for overlay measurement based on the analysis of the input data and the corresponding overlay measurements;
deploying the model for a wafer production run, wherein real-time inputs are obtained from the lithography process and the upstream processes and fed into the model;
generating a predicted overlay measurement using the model; and
adjusting the lithography process or the upstream processes such that the predicted overlay measurement correlates with an actual overlay measurement.

5. The method of claim 4 , further comprising:
normalizing the real-time inputs when a second statistical distribution of the real-time input has changed from a first statistical distribution of the input data.

6. The method of claim 5 , wherein the normalizing step is implemented by determining a z-score for the first and second statistical distributions.

7. The method of claim 4 , further comprising:
creating a virtual metrology model based on the data obtained from upstream processes; and
providing an output of the virtual metrology model as an input to the overlay measurement model.

8. The method of claim 4 , further comprising:
obtaining in-situ metrology data; and
providing the in-situ metrology data as an input to the overlay measurement model.

9. The method of claim 4 , further comprising:
performing a transformation of one or more sets of the input data; and
providing the transformed input data as an input to the overlay measurement model.

10. The method of claim 3 , wherein the data obtained from the lithography process and the upstream processes includes metrology and parametric data.

11. The method of claim 10 , wherein the metrology and parametric data from the lithography process includes feature critical dimensions, wafer shape, wafer geometry, film thickness, film resistivity, device channel length, device channel width, device channel depth, device operating thresholds, and device resistance.

12. The method of claim 10 , wherein the metrology and parametric data from the upstream processes includes, for each upstream process, process duration, process temperature, process pressure, process frequency, and optical measurements.

13. The method of claim 3 , wherein the overlay measurements are obtained using image-based overlay or diffraction-based overlay.

14. The method of claim 3 , wherein the analyzing step is performed by at least one machine learning algorithm.

15. The method of claim 3 , wherein the analyzing step is performed by a combination of machine learning algorithms.

16. The method of claim 3 , wherein the analyzing step is performed by a multi-step algorithm.

17. A non-transitory machine-readable medium having stored thereon one or more sequences of instructions, which instructions, when executed by one or more processors, cause the one or more processors to carry out the steps of:
obtaining a plurality of actual overlay measurements from a plurality of wafers in a plurality of production runs of a lithography process, wherein each actual overlay measurement indicates an offset between a first set of features formed on a first layer and a second set of features formed on a second layer above the first layer;
collecting a plurality of sets of input data, each set of input data collected from one of the plurality of production runs including data obtained from the lithography process and data obtained from at least one upstream process;
analyzing the sets of input data collected from each production run to determine a multi-variate relationship of the input data to the actual overlay measurements;
evaluating the multi-variate relationship to generate a predicted overlay measurement for each set of input data; and
adjusting the lithography process or the at least one upstream process such that the predicted overlay measurements correlate with the actual overlay measurements for each set of input data.

18. The non-transitory machine-readable medium of claim 17 , comprising further instructions that cause the one or more processors to carry out the steps of:
creating a model for overlay measurement based on the analysis of the input data and the corresponding overlay measurements;
deploying the model for a wafer production run, wherein real-time inputs are obtained from the lithography process and the upstream processes and fed into the model;
generating a predicted overlay measurement using the model; and
adjusting the lithography process or the upstream processes such that the predicted overlay measurement correlates with an actual overlay measurement.

19. A system, comprising:
at least one processor; and
a memory coupled to the processor comprising instructions executable by the processor, the instructions, when executed by the processor, cause the processor to:
obtain a plurality of actual overlay measurements from a plurality of wafers in a plurality of production runs of a lithography process, wherein each actual overlay measurement indicates an offset between a first set of features formed on a first layer and a second set of feat

Assignees

• Stream Mosaic Inc
• Pdf Solutions Inc

Inventors

• David Jeffrey Drue

Classifications

• H10P74/203

  Structural properties, e.g. testing or measuring thicknesses, line widths, warpage, bond strengths or physical defects · CPC title

• H10P74/207

  Electrical properties, e.g. testing or measuring of resistance, deep levels or capacitance-voltage characteristics · CPC title

• H10P72/0604

  Process monitoring, e.g. flow or thickness monitoring · CPC title

• H10P74/23 Primary

  characterised by multiple measurements, corrections, marking or sorting processes · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title