Guardbands in substrate processing systems
A method includes identifying trace data including a plurality of data points, the trace data being associated with production, via a substrate processing system, of substrates that have property values that meet threshold values. The method further includes determining, based on the trace data, a dynamic acceptable area outside of guardband limits. The method further includes causing, based on the dynamic acceptable area outside of the guardband limits, performance of a corrective action associated with the substrate processing system.
1. A method comprising:
identifying trace data comprising a plurality of data points, the trace data being associated with production, via a substrate processing system, of substrates that have property values that meet threshold values;
determining, based on the trace data, a dynamic acceptable area outside of guardband limits; and
causing, based on the dynamic acceptable area outside of the guardband limits, performance of a corrective action associated with the substrate processing system.
2. The method of claim 1 , wherein the trace data comprises a first portion and a second portion that occurs after the first portion, wherein the dynamic acceptable area is associated with one or more of drift or noise associated with the substrate processing system.
3. The method of claim 1 , wherein the determining of the dynamic acceptable area outside of the guardband limits comprises training a machine learning model with data input comprising the trace data to generate a trained machine learning model indicative of the dynamic acceptable area outside of the guardband limits.
4. The method of claim 1 further comprising adjusting the guardband limits based on incoming data that is within the dynamic acceptable area.
5. The method of claim 1 further comprising warping portions of the trace data horizontally without warping vertically to ignore out-of-phase factor while preserving vertical noise.
6. The method of claim 1 further comprising scaling portions of the trace data vertically and horizontally to ignore differing amplitudes and recipe end-pointing.
7. The method of claim 1 , wherein the dynamic acceptable area is a first value at a first portion of the guardband limits and a second value at a second portion of the guardband limits.
8. The method of claim 1 , wherein the trace data comprises historical trace data and simulated trace data, the simulated trace data being generated by applying one or more of drift, oscillation, noise, or spikes to the historical trace data.
9. The method of claim 1 , wherein the causing of the performance of the corrective action comprises:
providing additional trace data as input to a trained machine learning model;
receiving, from the trained machine learning model, output comprising predictive data; and
determining, based on the predictive data, that at least a portion of the additional trace data is outside of the dynamic acceptable area outside of the guardband limits.
10. A method comprising:
identifying trace data comprising a plurality of data points, the trace data being associated with substrate production via a substrate processing system;
comparing the trace data to an acceptable area outside of guardband limits; and
responsive to one or more data points of the trace data being within the acceptable area, updating the acceptable area outside of the guardband limits based on the trace data, wherein performance of a corrective action associated with the substrate processing system is based on at least a portion of the trace data being outside of the acceptable area outside of the guardband limits.
11. The method of claim 10 , wherein the trace data comprises a first portion and a second portion that occurs after the first portion, wherein the updating of the acceptable area is based on one or more of drift or noise associated with the substrate processing system.
12. The method of claim 10 , wherein the acceptable area outside of the guardband limits is associated with a trained machine learning model that was trained with data input comprising historical trace data over time.
13. The method of claim 12 , wherein the data input further comprises simulated trace data, the simulated trace data being generated by applying one or more of drift, oscillation, noise, or spikes to the historical trace data.
14. The method of claim 10 further comprising warping portions of the trace data horizontally without warping vertically to ignore out-of-phase factor while preserving vertical noise.
15. The method of claim 10 further comprising scaling portions of the trace data vertically and horizontally to ignore differing amplitudes and recipe end-pointing.
16. The method of claim 10 , wherein the acceptable area is a first value at a first portion of the guardband limits and is a second value at a second portion of the guardband limits.
17. The method of claim 10 , wherein the comparing of the trace data to the acceptable area outside of the guardband limits comprises:
providing the trace data as input to a trained machine learning model;
receiving, from the trained machine learning model, output comprising predictive data; and
determining, based on the predictive data, that at least a portion of the trace data is within the guardband limits, within the acceptable area, or outside of the acceptable area.
18. A non-transitory computer-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising:
identifying trace data comprising a plurality of data points, the trace data being associated with production, via a substrate processing system, of substrates that have property values that meet threshold values;
determining, based on the trace data, a dynamic acceptable area outside of guardband limits; and
causing, based on the dynamic acceptable area outside of the guardband limits, performance of a corrective action associated with the substrate processing system.
19. The non-transitory computer-readable storage medium of claim 18 , wherein the trace data comprises a first portion and a second portion that occurs after the first portion, wherein the dynamic acceptable area is associated with one or more of drift or noise associated with the substrate processing system.
20. The non-transitory computer-readable storage medium of claim 18 , wherein the determining of the dynamic acceptable area outside of the guardband limits comprises training a machine learning model with data input comprising the trace data to generate a trained machine learning model indicative of the dynamic acceptable area outside of the guardband limits.