Phase defect characterization using generative adversarial networks for extreme ultraviolet lithography

Hang Zheng; Sikun Li; Sikun Li; Wei Cheng; Wei Cheng; Shuai Yuan; Xiangzhao Wang

doi:10.1364/AO.480356

Applied Optics
Vol. 62,
Issue 5,
pp. 1243-1252
(2023)
•https://doi.org/10.1364/AO.480356

Phase defect characterization using generative adversarial networks for extreme ultraviolet lithography

Hang Zheng, Sikun Li, Wei Cheng, Shuai Yuan, and Xiangzhao Wang

Not Accessible

Your library or personal account may give you access

Get PDF
Email
Share
Get Citation
Copy Citation Text
Hang Zheng, Sikun Li, Wei Cheng, Shuai Yuan, and Xiangzhao Wang, "Phase defect characterization using generative adversarial networks for extreme ultraviolet lithography," Appl. Opt. 62, 1243-1252 (2023)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Check for updates

Related Topics
Table of Contents Category
- Fourier Optics and Optical Processing
Optics & Photonics Topics
?

The topics in this list come from the Optics and Photonics Topics applied to this article.

About this Article
History
- Original Manuscript: November 8, 2022
- Revised Manuscript: January 8, 2023
- Manuscript Accepted: January 8, 2023
- Published: February 6, 2023

Abstract

The multilayer defects of mask blanks in extreme ultraviolet (EUV) lithography may cause severe reflectivity deformation and phase shift. The profile information of a multilayer defect is the key factor for mask defect compensation or repair. This paper introduces an artificial neural network framework to reconstruct the profile parameters of multilayer defects in the EUV mask blanks. With the aerial images of the defective mask blanks obtained at different illumination angles and a series of generative adversarial networks, the method enables a way of multilayer defect characterization with high accuracy.

Full Article | PDF Article

More Like This

Extreme ultraviolet phase defect characterization based on complex amplitudes of the aerial images

Wei Cheng, Sikun Li, Xiangzhao Wang, and Zinan Zhang
Appl. Opt. 60(17) 5208-5219 (2021)

EUV multilayer defect characterization via cycle-consistent learning

Ying Chen, Yibo Lin, Rui Chen, Lisong Dong, Ruixuan Wu, Tianyang Gai, Le Ma, Yajuan Su, and Yayi Wei
Opt. Express 28(12) 18493-18506 (2020)

Repairing amplitude defects in multilayer-coated extreme-ultraviolet lithography reticles by use of a focused ion beam

Anton Barty, Stefan Hau-Riege, Dan Stearns, Miles Clift, Paul Mirkarimi, Eric Gullikson, Henry Chapman, and Don Sweeney
Appl. Opt. 43(36) 6545-6556 (2004)

Previous Article Next Article

Data availability

Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (11)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (7)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (6)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Reconstruction Part (Input i Image)
Layer	Filter	Output Size
Input	–	$95 \times 95 \times i$
Convolution	$7 \times 7 N = 16$	$95 \times 95 \times 16$
CM	$3 \times 3 N = 32$	$48 \times 48 \times 32$
CM	$3 \times 3 N = 64$	$24 \times 24 \times 64$
CM	$3 \times 3 N = 128$	$12 \times 12 \times 128$
DM	$3 \times 3 N = 256 D i l a t i o n = 2$	$12 \times 12 \times 256$
DM	$3 \times 3 N = 512 D i l a t i o n = 4$	$12 \times 12 \times 512$
Convolution	$3 \times 3 N = 512 D i l a t i o n = 2$	$12 \times 12 \times 512$
Convolution	$3 \times 3 N = 512$	$12 \times 12 \times 512$
Max-pool	$3 \times 3$	$4 \times 4 \times 512$
Flatten	–	$8192 \times 1$
Dense 100	–	$100 \times 1$
Dense 50	–	$50 \times 1$
Output	–	$4 \times 1$
Discriminator Part ( $n$ Times in a Loop)
Layer	Filter	Output Size
Input	–	$4 \times 1$
Dense 50	–	$50 \times 1$
Dense 100	–	$100 \times 1$
Dense $512 \times 12 \times 12$	–	$73728 \times 1$
Dense $1 \times 95 \times 95$	–	$9025 \times 1$
Output	Reshape	$95 \times 95 \times 1$

Illumination Angle	$l_{1}$	$l_{2}$	$l_{3}$	$l_{4}$	$l_{5}$	$l_{6}$	$l_{7}$	$l_{8}$	$l_{9}$
$θ_{r}$ ( $^{\circ}$ )	0	2.03	2.84	2.03	2.84	2.03	2.84	2.03	2.84
$φ$ ( $^{\circ}$ )	0	0	45	90	135	180	−135	−90	−45

	Number	$h_{t o p}$	$ω_{t o p}$	$h_{b o t}$	$ω_{b o t}$	Average
	of Input	ARE	ARE	ARE	ARE	ARE
Model	Images ( $i$ )	(%)	(%)	(%)	(%)	(%)
DRN ( $i = 1$ )	1	4.51	1.82	2.17	1.23	2.43
DRN ( $i = 3$ )	3	4.80	0.95	0.83	0.89	1.86
DRN ( $i = 5$ )	5	3.29	1.06	1.13	1.32	1.70
DRN ( $i = 7$ )	7	3.41	1.41	1.22	1.58	1.91
DRN ( $i = 9$ )	9	4.70	1.03	1.16	0.96	1.96

		$h_{t o p}$	$ω_{t o p}$	$h_{b o t}$	$ω_{b o t}$	Average
		ARE	ARE	ARE	ARE	ARE
Model	$n$	(%)	(%)	(%)	(%)	(%)
DRN $+$ GAN ( $i = 5$ )	5	1.74	0.83	0.83	0.80	1.05
DRN $+$ GAN ( $i = 5$ )	2	1.61	0.25	1.18	0.82	0.97
DRN $+$ GAN ( $i = 5$ )	1	2.13	0.25	1.35	1.61	1.34
DRN ( $i = 5$ )	0	3.29	1.06	1.13	1.32	1.70

		$h_{t o p}$	$ω_{t o p}$	$h_{b o t}$	$ω_{b o t}$	Average
Defect		ARE	ARE	ARE	ARE	ARE
Type	Model	(%)	(%)	(%)	(%)	(%)
Bump	FPI $+$ ANN	2.75	0.47	3.49	1.75	2.12
Pit		3.71	0.24	2.32	1.28	1.89
Bump	DRN ( $i = 5$ )	3.17	0.22	2.74	1.15	1.82
Pit		3.29	1.06	1.13	1.32	1.70
Bump	DRN $+$ GAN	2.96	0.28	2.38	1.18	1.70
Pit	( $i = 5$ , $n = 5$ )	1.74	0.83	0.83	0.80	1.05
Bump	DRN $+$ GAN	2.47	0.39	2.19	0.98	1.51
Pit	( $i = 5$ , $n = 2$ )	1.61	0.25	1.18	0.82	0.97
Bump	DRN $+$ GAN	1.97	0.36	2.60	1.25	1.54
Pit	( $i = 9$ , $n = 2$ )	2.12	0.24	1.06	0.93	1.10

				$h_{t o p}$	$ω_{t o p}$	$h_{b o t}$	$ω_{b o t}$	Average
				ARE	ARE	ARE	ARE	ARE
Defect Type	Model	Noises Type	SNR (dB)	(%)	(%)	(%)	(%)	(%)
Bump	DRN $+$ GAN ( $i = 5$ , $n = 2$ )	without	/	2.47	0.39	2.19	0.98	1.51
Bump		Gaussian	37.2	4.05	0.96	16.01	6.46	6.87
Bump		Gaussian	42.2	3.96	0.53	8.24	2.89	3.91
Bump		Gaussian	47.2	2.66	0.42	4.80	1.88	2.44
Bump		Poisson	37.7	2.90	0.81	11.78	6.52	5.48
Bump	FPI $+$ ANN	Gaussian	47.2	2.92	0.51	6.94	2.97	3.33
Bump	FPI $+$ ANN	Poisson	37.7	5.22	0.96	10.61	5.93	5.68
Pit	DRN $+$ GAN ( $i = 5$ , $n = 2$ )	without	/	1.61	0.25	1.18	0.82	0.97
Pit		Gaussian	37.2	2.53	0.51	3.59	2.08	2.18
Pit		Gaussian	42.2	2.41	0.45	2.13	1.73	1.68
Pit		Gaussian	47.2	2.37	0.38	1.40	1.04	1.30
Pit		Poisson	37.7	2.26	0.51	2.95	2.40	2.03
Pit	FPI $+$ ANN	Gaussian	47.2	3.62	0.39	2.89	1.21	2.03
Pit	FPI $+$ ANN	Poisson	37.7	3.82	0.50	4.58	2.19	2.77

Abstract

Data availability

Cited By

Figures (11)

Tables (7)

Equations (6)

Applied Optics