Determining a correction to a process
A method for configuring a semiconductor manufacturing process, the method including: obtaining a first value of a first parameter based on measurements associated with a first operation of a process step in the semiconductor manufacturing process and a first sampling scheme; using a recurrent neural network to determine a predicted value of the first parameter based on the first value; and using the predicted value of the first parameter in configuring a subsequent operation of the process step in the semiconductor manufacturing process.
1. A method comprising:
obtaining a set of values of a parameter associated with a plurality of positions on a substrate;
obtaining a first matrix of values based on evaluation of one or more base functions at the plurality of positions;
obtaining a second matrix of values based on training, by a hardware computer, a matrix of adaptable numbers to previously obtained sets of values of the parameter associated with previous substrates;
using the first matrix and second matrix of values and the obtained sets of values of the parameter to determine coefficients of a model; and
using the coefficients and second matrix of values to provide modeled values of the parameter.
2. The method according to claim 1 , wherein the parameter is overlay.
3. The method according to claim 1 , wherein the plurality of positions is associated with a sparse measurement scheme across one or more substrates being subject to a semiconductor manufacturing process.
4. The method according to claim 1 , wherein the base functions are polynomials.
5. The method according to claim 1 , wherein the base functions are mutually orthogonal.
6. The method according to claim 1 , wherein the plurality of positions relate to an exposure field on the substrate.
7. The method according to claim 1 , wherein the adaptable numbers are freeform model parameters.
8. The method according to claim 1 , wherein the previously obtained sets of values of the parameter are associated with densely measured previous substrates.
9. The method according to claim 1 , wherein the training is based on using a machine learning model to establish relations between the adaptable numbers and a fingerprint of the parameter across each individual substrate out of the previous substrates.
10. The method according to claim 9 , wherein the machine learning model is based on a neural network.
11. A method comprising:
obtaining a first set of measured values of a patterning process parameter associated with a plurality of positions on a substrate; and
using, by a hardware computer system, the first set of measured values as an input to a machine learning model trained to map one or more measured values of the parameter to one or more predicted values of the parameter to obtain a second set of modeled values of the parameter.
12. The method according to claim 11 , wherein the machine learning model is a neural network.
13. The method according to claim 11 , wherein the machine learning model is trained to optimize the similarity between the one or more measured values of the parameter and the one or more predicted values of the parameter.
14. A computer program product comprising a non-transitory computer-readable medium comprising instructions therein, the instructions, upon execution by a computer system, configured to cause the computer system to at least:
obtain a set of values of a parameter associated with a plurality of positions on a substrate;
obtain a first matrix of values based on evaluation of one or more base functions at the plurality of positions;
obtain a second matrix of values based on training a matrix of adaptable numbers to previously obtained sets of values of the parameter associated with previous substrates;
use the first matrix and second matrix of values and the obtained sets of values of the parameter to determine coefficients of a model; and
use the coefficients and second matrix of values to provide modeled values of the parameter.
15. The computer program product of claim 14 , wherein the base functions are polynomials.
16. The computer program product of claim 14 , wherein the adaptable numbers are freeform model parameters.
17. The computer program product of claim 14 , wherein the machine learning model is based on a neural network.
18. A computer program product comprising a non-transitory computer-readable medium comprising instructions therein, the instructions, upon execution by a computer system, configured to cause the computer system to at least:
obtain a first set of measured values of a patterning process parameter associated with a plurality of positions on a substrate; and
use the first set of measured values as an input to a machine learning model trained to map one or more measured values of the parameter to one or more predicted values of the parameter to obtain a second set of modeled values of the parameter.
19. The computer program product of claim 18 , wherein the machine learning model is a neural network.
20. The computer program product of claim 18 , wherein the machine learning model is trained to optimize the similarity between the one or more measured values of the parameter and the one or more predicted values of the parameter.