System and method to emulate finite element model based prediction of in-plane distortions due to semiconductor wafer chucking

Patent metadata

Field	Value
Publication number	US-10025894-B2
Application number	US-201615172667-A
Country	US
Kind code	B2
Filing date	Jun 3, 2016
Priority date	Oct 11, 2012
Publication date	Jul 17, 2018
Grant date	Jul 17, 2018

Abstract

Systems and methods for prediction of in-plane distortions (IPD) due to wafer shape in semiconductor wafer chucking process is disclosed. A series of Zernike basis wafer shapes process to emulate the non-linear finite element (FE) contact mechanics model based IPD prediction is utilized in accordance with one embodiment of the present disclosure. The emulated FE model based prediction process is substantially more efficient and provides accuracy comparable to the FE model based IPD prediction that utilizes full-scale 3-D wafer and chuck geometry information and requires computation intensive simulations. Furthermore, an enhanced HOS IPD/OPD prediction process based on a series of Zernike basis wafer shape images is also disclosed.

First claim

What is claimed is:

1. A computer implemented method for providing in-plane distortion (IPD) prediction, the method comprising:
(a) generating, with one or more processors, a series of Zernike basis wafer shapes;
(b) performing, with the one or more processors, finite element (FE) model based IPD prediction for the series of Zernike basis wafer shapes;
(c) performing, with the one or more processors, higher order shape (HOS) based IPD prediction for the series of Zernike basis wafer shapes;
(d) comparing, with the one or more processors, the FE model based IPD prediction and the HOS based IPD prediction of a particular Zernike basis wafer shape of the series of Zernike basis wafer shapes to determine whether said particular Zernike basis wafer shape produces large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction;
(e) storing, in memory, the Zernike basis wafer shapes that produce large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction; and
(f) configuring a process tool to compensate for at least one of overlay error or misalignment based on a HOS based IPD prediction for a given wafer utilizing the stored Zernike basis wafer shapes.

2. The method of claim 1 , wherein providing the HOS based IPD prediction for the given wafer utilizing the stored Zernike basis wafer shapes further comprises:
obtaining a wafer shape image of the given wafer;
performing HOS based IPD prediction for the wafer shape image of the given wafer;
decomposing the wafer shape image to a linear combination of Zernike basis images;
identifying, among the linear combination of Zernike basis images, a subset of Zernike basis images that produce large prediction differences, wherein the subset of Zernike basis images that produce large prediction differences are identified based on the stored Zernike basis wafer shapes; and
combining the FE model based IPD prediction for each one of the subset of Zernike basis images that produce large prediction differences with the HOS based IPD prediction result for the given wafer.

3. The method of claim 1 , wherein a particular Zernike basis wafer shape is identified as a Zernike basis wafer shape that produces large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction when the prediction differences exceed a predetermined threshold.

4. The method of claim 1 , wherein steps (a) through (e) are a part of a prediction process and are performed prior to and independently from step (f).

5. The method of claim 1 , wherein step (f) is performed prior to and after a wafer patterning process for at least one of: overlay error control or alignment control.

6. The method of claim 1 , wherein the predicted IPD for the given wafer is utilized as control input for controlling at least one of: a Rapid Thermal Processing (RTP) process, a Chemical-Mechanical Planarization (CMP) process, or a Chemical Vapor Deposition (CVD) process.

7. The method of claim 1 , further comprising:
selectively masking off at least one non-functional space on at least one of: a front, a back and a shape image map of the given wafer.

8. A system for providing in-plane distortion (IPD) prediction for a given wafer, the system comprising:
an optical system configured for obtaining a wafer shape of the given wafer; and
an IPD prediction module in communication with the optical system, the IPD prediction module configured for:
(a) generating a series of Zernike basis wafer shapes;
(b) performing finite element (FE) model based IPD prediction for the series of Zernike basis wafer shapes;
(c) performing higher order shape (HOS) based IPD prediction for the series of Zernike basis wafer shapes;
(d) comparing the FE model based IPD prediction and the HOS based IPD prediction of a particular Zernike basis wafer shape of the series of Zernike basis wafer shapes to determine whether said particular Zernike basis wafer shape produces large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction;
(e) storing the Zernike basis wafer shapes that produce large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction; and

(f) configuring a process tool to compensate for at least one of overlay error or misalignment based on a HOS based IPD prediction for a given wafer utilizing the stored Zernike basis wafer shapes.

9. The system of claim 8 , wherein providing the HOS based IPD prediction for the given wafer utilizing the stored Zernike basis wafer shapes further comprises:
obtaining a wafer shape image of the given wafer;
performing HOS based IPD prediction for the wafer shape image of the given wafer;
decomposing the wafer shape image to a linear combination of Zernike basis images;
identifying, among the linear combination of Zernike basis images, a subset of Zernike basis images that produce large prediction differences, wherein the subset of Zernike basis images that produce large prediction differences are identified based on the stored Zernike basis wafer shapes; and
combining the FE model based IPD prediction for each one of the subset of Zernike basis images that produce large prediction differences to the HOS based IPD prediction result for the given wafer.

10. The system of claim 8 , wherein a particular Zernike basis wafer shape is identified as a Zernike basis wafer shape that produces large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction when the prediction differences exceed a predetermined threshold.

11. The system of claim 8 , wherein the IPD prediction module performs steps (a) through (e) as a part of a prediction process and performs step (f) independently from steps (a) through (e).

12. The system of claim 8 , wherein the IPD prediction module performs step (f) prior to and after a wafer patterning process for at least one of: overlay error control or alignment control.

13. The system of claim 8 , wherein the predicted IPD for the given wafer is utilized as control input to a downstream application, and wherein the downstream application includes at least one of: a Rapid Thermal Processing (RTP) process, a Chemical-Mechanical Planarization (CMP) process, or a Chemical Vapor Deposition (CVD) process.

14. The system of claim 8 , wherein the IPD prediction module is further configured for selectively masking off at least one non-functional space on at least one of: a front, a back and a shape image map of the given wafer.

15. A computer implemented method for overlay error prediction, the method comprising:
(a) generating, with one or more processors, a series of Zernike basis wafer shapes;
(b) performing, with the one or more processors, finite element (FE) model based IPD prediction for the series of Zernike basis wafer shapes;
(c) performing, with the one or more processors, higher order shape (HOS) based IPD prediction for the series of Zernike basis wafer shapes;
(d) comparing, with the one or more processors, the FE model based IPD prediction and the HOS based IPD prediction of a particular Zernike basis wafer shape of the series of Zernike basis wafer shapes to determine whether said particular Zernike basis wafer shape produces large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction;
(e) storing, in memory, the Zernike basis wafer shapes that produce large prediction differences between the FE model based IPD prediction and the HOS based IPD prediction;
(f) performing, with the one or more

Assignees

• Kla Tencor Corp

Inventors

• Vukkadala Pradeep
• Veeraraghavan Sathish
• Sinha Jaydeep
• Chen Haiguang
• Kirk Michael

Classifications

• H10P72/0616

  Monitoring of warpages, curvatures, damages, defects or the like · CPC title

• G06F30/23 Primary

  using finite element methods [FEM] or finite difference methods [FDM] · CPC title

• H01L21/67288

  Electricity · mapped topic

• G06F17/5018 Primary

  Physics · mapped topic