Thickness estimation method and processing control method

Patent metadata

Field	Value
Publication number	US-11668558-B2
Application number	US-202117214830-A
Country	US
Kind code	B2
Filing date	Mar 27, 2021
Priority date	Aug 28, 2020
Publication date	Jun 6, 2023
Grant date	Jun 6, 2023

Abstract

A thickness estimation method may include: obtaining a test spectrum image; obtaining test spectrum data; measuring a thickness of a test layer formed on the test substrate at the plurality of positions; generating a regression analysis model using a correlation between the thickness of the test layer and the test spectrum data; obtaining a spectrum image; and estimating a thickness of a target layer over the entire area of the semiconductor substrate by applying the spectrum image to the regression analysis model. The thickness corresponding to the entire area of the semiconductor substrate that is being transferred is estimated using the thickness estimation method according to an exemplary embodiment in the present disclosure, such that whether or not processing is normally performed may be examined without requiring a separate time. In addition, an examination result may be feedbacked to processing equipment to improve production yield.

First claim

What is claimed is:

1. A thickness estimation method comprising:
obtaining a test spectrum image from reflected light obtained by irradiating a test substrate with light;
obtaining test spectrum data included in a predetermined wavelength band at a plurality of positions on the test substrate, from the test spectrum image;
measuring a thickness of a test layer formed on the test substrate at the plurality of positions;
generating a regression analysis model using a correlation between the thickness of the test layer and the test spectrum data at the plurality of positions, and storing the regression analysis model in a memory;
irradiating a semiconductor substrate that is being transferred with light;
obtaining a spectrum image corresponding to an entire area of the semiconductor substrate; and
estimating a thickness of a target layer corresponding to the test layer over the entire area of the semiconductor substrate by applying the spectrum image to the regression analysis model.

2. The thickness estimation method of claim 1 ,
wherein the spectrum image is obtained by an optical system, and
wherein the optical system includes:
a lens on which light reflected from the semiconductor substrate is incident;
a spectrometer module splitting the incident light; and
a scan camera capturing an image of the split light in a form of a spectrum.

3. The thickness estimation method of claim 2 ,
wherein the scan camera includes a first camera capturing an image of a first region of the semiconductor substrate and a second camera capturing an image of a second region of the semiconductor substrate, and
wherein the images of the first region and the second region are combined to form the spectrum image.

4. The thickness estimation method of claim 1 , wherein the obtaining the spectrum image includes:
obtaining line scan spectrum images using an optical system;
generating an original spectrum image of the semiconductor substrate using wavelength data included in the line scan spectrum images; and
generating a corrected spectrum image by applying a circle correction algorithm to the original spectrum image.

5. The thickness estimation method of claim 1 , wherein the thickness of the test layer at the plurality of positions on the test substrate is measured using optical critical dimension (OCD) measurement.

6. The thickness estimation method of claim 1 , further comprising:
performing normalization of the test spectrum data after obtaining the test spectrum image.

7. The thickness estimation method of claim 1 , wherein the predetermined wavelength band is determined based on quality of at least one of the test layer and the target layer.

8. The thickness estimation method of claim 1 , wherein the regression analysis model is automatically optimized through periodic update of the test spectrum data.

9. The thickness estimation method of claim 1 ,
wherein the semiconductor substrate is irradiated with the light by an illumination unit disposed on a transfer path of the semiconductor substrate, and
wherein the spectrum image is obtained using an optical system that is integrated with the illumination unit.

10. The thickness estimation method of claim 1 , wherein the obtaining the spectrum image is performed by an optical system while the semiconductor substrate is transferred from a first device to a second device by a transfer robot.

11. The thickness estimation method of claim 10 , further comprising:
performing a semiconductor processing before the semiconductor substrate is transferred from the first device to the second device,
wherein the estimating the thickness of the target layer over the entire area of the semiconductor substrate is performed separately from the semiconductor processing.

12. The thickness estimation method of claim 10 , wherein when the transfer robot transfers the semiconductor substrate, a transfer speed of the semiconductor substrate changes over time.

13. A processing control method comprising:
obtaining a test spectrum image from reflected light obtained by irradiating a test substrate with light;
generating a regression analysis model using a correlation between test spectrum data included in the test spectrum image and thickness data of a test layer formed on the test substrate, and storing the regression analysis model in a memory;
performing a semiconductor processing on a first semiconductor substrate to which a first processing control parameter is applied;
obtaining a spectrum image of the first semiconductor substrate on a transfer path of the first semiconductor substrate subjected to the semiconductor processing;
obtaining spectrum data included in a predetermined wavelength band from the spectrum image;
updating the regression analysis model using a correlation between the spectrum data and the first processing control parameter;
determining a second processing control parameter using the regression analysis model; and
performing the semiconductor processing on a second semiconductor substrate, which is the same as the first semiconductor substrate, and to which the second processing control parameter is applied.

14. The processing control method of claim 13 ,
wherein the semiconductor processing is chemical-mechanical polishing (CMP) processing, and
wherein the first processing control parameter and the second processing control parameter include a processing pressure and a processing time.

15. The processing control method of claim 13 ,
wherein the semiconductor processing is spin-coating processing, and
wherein the first processing control parameter and the second processing control parameter include a baking time and a baking temperature.

16. The processing control method of claim 13 , wherein the second processing control parameter is determined by setting a target thickness of a target layer formed on the second semiconductor substrate by the semiconductor processing, and applying the target thickness to the regression analysis model.

17. The processing control method of claim 13 , wherein the second processing control parameter is automatically optimized, and is determined by the regression analysis model that is updated using the correlation between the spectrum data and the first processing control parameter.

18. A thickness estimation method comprising:
obtaining a test spectrum image corresponding to a plurality of positions on a test substrate;
generating a regression analysis model using a correlation between test spectrum data included in the test spectrum image and thickness data of a test layer formed on the test substrate, and storing the regression analysis model in a memory;
irradiating a semiconductor substrate that is being transferred with light and obtaining a spectrum image corresponding to an entire area of the semiconductor substrate; and
estimating a thickness of a target layer over the entire area of the semiconductor substrate by applying spectrum data based on the spectrum image of the semiconductor substrate to the regression analysis model.

19. The thickness estimation method of claim 18 ,
wherein the regression analysis model is continuously updated using a correlation between thickness data of the target layer and a processing control parameter with which processing for the semiconductor substrate is performed, and
wherein the processing is performed by adjusting the processing control parameter for the semiconductor substrate.

20. The thickness estimation method of claim 18 ,
wherein the regressio

Assignees

• Samsung Electronics Co Ltd

Inventors

• Kim Jongsu
• Park Wansung
• Cho Doohyun
• Kim Sungha
• Wi Jaeyoun
• Hong Kijoo
• Kim Taejoong
• Ryu Youngsu
• LEE Kwangsung
• Hong Min

Classifications

• G01N21/9501 Primary

  Semiconductor wafers (manufacturing processes per se of semiconductor devices implementing a measuring step H10P74/20) · CPC title

• G01N21/8422

  Investigating thin films, e.g. matrix isolation method · CPC title

• G01B11/0633 Primary

  using one or more discrete wavelengths · CPC title

• G01B11/06

  for measuring thickness {; e.g. of sheet material (thickness measurement by thermal means G01B21/085)} · CPC title

• G01B2210/56

  Measuring geometric parameters of semiconductor structures, e.g. profile, critical dimensions or trench depth · CPC title