Spectrographic monitoring using a neural network

Patent metadata

Field	Value
Publication number	US-11966212-B2
Application number	US-202016984001-A
Country	US
Kind code	B2
Filing date	Aug 3, 2020
Priority date	Nov 30, 2016
Publication date	Apr 23, 2024
Grant date	Apr 23, 2024

Abstract

A method of processing a substrate includes subjecting a substrate to processing that modifies a thickness of an outer layer of the substrate, measuring a spectrum of light reflected from the substrate during processing, reducing the dimensionality of the measured spectrum to generate a plurality of component values, generating a characterizing value using an artificial neural network, and determining at least one of whether to halt processing of the substrate or an adjustment for a processing parameter based on the characterizing value. The artificial neural network has a plurality of input nodes to receive the plurality of component values, an output node to output the characterizing value, and a plurality of hidden nodes connecting the input nodes to the output node.

First claim

What is claimed is:

1. A computer program product for controlling processing of a substrate, the computer program product tangibly embodied in a non-transitory computer readable media and comprising instructions for causing a one or more computers to:
receive, from an in-situ optical monitoring system, a measured spectrum of light reflected from a substrate undergoing processing that modifies a thickness of an outer layer of the substrate;
reduce a dimensionality of the measured spectrum to generate a plurality of component values by performing, on the measured spectrum, a transformation based on a plurality of reference spectra;
receive one or more values of one or more processing or environmental variables other than the plurality of component values;
for each respective scan of a plurality of scans, generate a value for the thickness for the respective scan of the outer layer of the substrate undergoing processing using an artificial neural network in an inference mode, the value for the thickness representing a thickness for the respective scan of a sensor of the in-situ optical monitoring system to measure the measured spectrum of light, wherein the artificial neural network comprises
a plurality of input nodes, the plurality of input nodes including a multiplicity of input nodes to receive in the inference mode the plurality of component values and one or more input nodes of the plurality of input nodes to receive in the inference mode the values of the one or more processing or environmental variables other than the plurality of component values,
an output node to output the value for the thickness of the outer layer of the substrate for the respective scan of the sensor of the in-situ optical monitoring system to measure the measured spectrum of light, and
a plurality of hidden nodes connecting the plurality of input nodes to the output node; and

determine at least one of whether to halt processing of the substrate or an adjustment for a processing parameter based on the value for the thickness of the outer layer of the substrate undergoing processing.

2. The computer program product of claim 1 , wherein the transformation is based on spectral feature components extracted from the plurality of reference spectra by feature extraction applied to the plurality of reference spectra.

3. The computer program product of claim 2 , comprising instructions to perform feature extraction on the plurality of reference spectra to generate the spectral feature components.

4. The computer program product of claim 3 , wherein the instructions to perform feature extraction comprise instructions to perform principal component analysis, singular value decomposition, independent component analysis, or autoencoding.

5. The computer program product of claim 1 , comprising instructions to perform dimensional reduction on two or more of the plurality of reference spectra that have known characteristic values to generate training data.

6. The computer program product of claim 5 , comprising instructions to train the artificial neural network by backpropagation using the training data and the known characteristic values.

7. The computer program product of claim 5 , wherein the two or more of the plurality of reference spectra includes fewer spectra than the plurality of reference spectra.

8. The computer program product of claim 1 , wherein the one or more processing or environmental variables other than the plurality of component values include at least one of a measurement from another sensor in a processing system in which the substrate undergoes the processing, a value from a processing recipe stored by a controller of the processing system, or a value of a variable tracked by the controller.

9. The computer program product of claim 1 , comprising instructions to cause a chemical mechanical polishing system to polish the substrate, and to receive the measured spectrum of light from the in-situ optical monitoring system during polishing of the substrate.

10. The computer program product of claim 9 , wherein the instructions to determine at least one of whether to halt processing of the substrate or an adjustment for a processing parameter comprise instructions to modify a pressure in a carrier head of the chemical mechanical polishing system.

11. The computer program product of claim 8 , wherein the measurement from the another sensor in the processing system in which the substrate undergoes the processing comprises a measurement of a temperature of the substrate, and wherein the measurement is obtained from a temperature sensor.

12. The computer program product of claim 8 , wherein the value from the processing recipe stored by the controller of the processing system comprises a polishing parameter used for polishing the substrate, and wherein the value is obtained from the controller.

13. A method of processing a substrate, comprising:
subjecting a substrate to processing that modifies a thickness of an outer layer of the substrate;
measuring during the processing with an in-situ optical monitoring system a measured spectrum of light reflected from the substrate undergoing processing;
reducing a dimensionality of the measured spectrum to generate a plurality of component values by performing, on the measured spectrum, a transformation based on a plurality of reference spectra;
receiving one or more values of one or more processing or environmental variables other than the plurality of component values;
for each scan of a plurality of scans, generating a value for the thickness for each scan of the outer layer of the substrate undergoing processing using an artificial neural network in an inference mode the value for the thickness representing a thickness for each scan of a sensor of the in-situ optical monitoring system to measure the measured spectrum of light, wherein the artificial neural network comprises
a plurality of input nodes, the plurality of input nodes including a multiplicity of input nodes to receive in the inference mode the plurality of component values and one or more input nodes of the plurality of input nodes to receive in the inference mode the values of the one or more processing or environmental variables other than the plurality of component values,
an output node to output the value for the thickness of the outer layer of the substrate for each scan of the sensor of the in-situ optical monitoring system to measure the measured spectrum of light, and
a plurality of hidden nodes connecting the plurality of input nodes to the output node; and

determining at least one of whether to halt processing of the substrate or an adjustment for a processing parameter based on the value for the thickness of the outer layer of the substrate undergoing processing.

14. The method of claim 13 , wherein the processing comprises chemical mechanical polishing.

15. The method of claim 13 , comprising performing feature extraction on the plurality of reference spectra to generate the transformation.

16. The method of claim 15 , wherein the feature extraction comprises principal component analysis, singular value decomposition, independent component analysis, or autoencoding.

17. The method of claim 14 , comprising performing dimensional reduction on two or more of the plurality of reference spectra that have known characteristic values to generate training data.

18. The method of claim 17 , comprising training the artificial neural network by backpropagation using the training data and the known characteristic values.

19. A polishing system,

Assignees

• Applied Materials Inc

Inventors

• Cherian Benjamin

Classifications

• H10P74/23

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

• H10P72/0428

  Apparatus for mechanical treatment or grinding or cutting · CPC title

• H10P52/00

  Grinding, lapping or polishing of wafers, substrates or parts of devices · CPC title

• H10P74/203 Primary

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

• G06N3/0499

  Feedforward networks · CPC title