Method and apparatus for determining etch process parameters

Patent metadata

Field	Value
Publication number	US-10784174-B2
Application number	US-201715783301-A
Country	US
Kind code	B2
Filing date	Oct 13, 2017
Priority date	Oct 13, 2017
Publication date	Sep 22, 2020
Grant date	Sep 22, 2020

Abstract

A method for processing a substrate in a processing chamber using at least one time trace based prediction model is provided. A substrate is dry processed, where the dry processing creates at least one gas by-product. A concentration of the at least one gas by-product is measured. A time trace of the concentration of the at least one gas by-product is determined. The determined time trace of the concentration is provided as input for the at least one time trace based prediction model to obtain at least one process output. The at least one process output is used to adjust at least one process parameter.

First claim

What is claimed is:

1. A method for processing a substrate in a processing chamber using at least one time trace based prediction model, comprising:
dry processing a substrate, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product, wherein the concentration of the at least one gas by-product is measured based on light absorption, wherein the concentration of the at least one gas by-product is measured using a multi-pass gas cell with an optical path length of at least one meter, wherein the at least one gas by-product is measured after the at least one gas by-product has passed through an exhaust pump, by placing the multi-pass gas cell on an output side of the exhaust pump;
obtaining a time trace of the concentration of the at least one gas by-product;
providing the obtained time trace of the concentration as input for the at least one time trace based prediction model to obtain at least one process output; and
using the at least one process output to adjust at least one process parameter, wherein the using the at least one process output to adjust at least one process parameter uses the at least one process output to adjust the at least one process parameter in real time.

2. The method, as recited in claim 1 , wherein the at least one time trace based prediction model is based on a multi-variate method.

3. The method, as recited in claim 1 , further comprising calculating a lump sum of absorption values from the obtained time trace of the concentration.

4. The method, as recited in claim 1 , wherein the at least one time trace based prediction model is further based on at least one of etch rate, uniformity, critical dimension, etch profile, wet clean, etch performance, chamber matching, and chamber cleaning performance.

5. The method, as recited in claim 1 , wherein the using the at least one process output comprises determining a change in at least one of etch rate, uniformity, critical dimension, and etch profile.

6. The method, as recited in claim 1 , wherein the using the at least one process output comprises determining if at least one of etch rate, uniformity, CD, and etch profile is out of specification.

7. A method for processing a substrate in a processing chamber using at least one time trace based prediction model, comprising:
dry processing a substrate, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product, wherein the concentration of the at least one gas by-product is measured based on light absorption, wherein the concentration of the at least one gas by-product is measured using a multi-pass gas cell with an optical path length of at least one meter, wherein the at least one gas by-product is measured after the at least one gas by-product has passed through an exhaust pump, by placing the multi-pass gas cell on an output side of the exhaust pump;
obtaining a time trace of the concentration of the at least one gas by-product;
providing the obtained time trace of the concentration as input for the at least one time trace based prediction model to obtain at least one process output, wherein the at least one process output includes at least two process outputs; and
using the at least one process output to adjust at least one process parameter.

8. A method for processing a substrate in a processing chamber using at least one time trace based prediction model, comprising:
dry processing a substrate, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product;
obtaining a time trace of the concentration of the at least one gas by-product;
providing the obtained time trace of the concentration as input for the at least one time trace based prediction model to obtain at least one process output; and
using the at least one process output to adjust at least one process parameter, wherein the using the at least one process output to adjust at least one process parameter, comprises using the time trace based prediction model to determine an upper process limit and a lower process limit and comparing the at least one process output to the upper process limit and the lower process limit.

9. A method for processing a substrate in a processing chamber using at least one time trace based prediction model, comprising:
dry processing a substrate, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product, wherein the concentration of the at least one gas by-product is measured based on light absorption, wherein the concentration of the at least one gas by-product is measured using a multi-pass gas cell with an optical path length of at least one meter, wherein the at least one gas by-product is measured after the at least one gas by-product has passed through an exhaust pump, by placing the multi-pass gas cell on an output side of the exhaust pump, wherein the measuring the concentration of the at least one gas by-product measures the at least one gas by-product with an accuracy of parts per trillion;
obtaining a time trace of the concentration of the at least one gas by-product;
providing the obtained time trace of the concentration as input for the at least one time trace based prediction model to obtain at least one process output; and
using the at least one process output to adjust at least one process parameter.

10. A method of creating a time trace based prediction model, comprising:
dry processing a plurality of substrates using process parameters, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product, wherein the concentration of the at least one gas by-product is measured based on light absorption, wherein the concentration of the at least one gas by-product is measured using a multi-pass gas cell with an optical path length of at least one meter, wherein the at least one gas by-product is measured after the at least one gas by-product has passed through an exhaust pump, by placing the multi-pass gas cell on an output side of the exhaust pump, wherein the measuring the concentration of the at least one gas by-product measures the at least one gas by-product with an accuracy of parts per trillion;
recording the process parameters;
obtaining a plurality of time traces of the concentration of the at least one gas by-product versus time for each substrate;
measuring output parameters; and
using the process parameters, obtained plurality of time traces, and measured output parameters to create a time trace based prediction model of at least one process output.

11. The method, as recited in claim 10 , wherein the at least one gas by-product comprises a silicon containing gas.

12. The method, as recited in claim 10 , wherein the using the process parameters, obtained plurality of time traces, and measured output parameters to create the time trace based prediction model of at least one process uses at least one of multivariable analysis or neural networking.

13. The method, as recited in claim 10 , wherein the at least one process output includes at least one of etch rate, uniformity, critical dimension, etch profile, wet clean, etch performance, chamber matching, and chamber cleaning performance.

14. A method of creating a time trace based prediction model, comprising:
dry processing a plurality of substrates using process parameters, wherein the dry processing creates at least one gas by-product;
measuring a concentration of the at least one gas by-product, wherein the conc

Assignees

• Lam Res Corp

Inventors

• Kabouzi Yassine
• Albarede Luc

Classifications

• H10P50/283

  by chemical means · CPC title

• H10P50/268

  of silicon-containing layers · CPC title

• H10P50/242 Primary

  of Group IV materials · CPC title

• H10P72/0604

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

• H10P72/0421

  for drying etching · CPC title