Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography
Wafer chips are manufactured in the semiconductor industry through various process technologies. Photolithography is one of these processes, aligning the wafer and scanning the circuit pattern on the wafer on which the photoresist film is formed by irradiating light onto the circuit pattern drawn on...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-03-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/5/2721 |
| _version_ | 1827651777450737664 |
|---|---|
| author | Juyong Park Jongpil Jeong |
| author_facet | Juyong Park Jongpil Jeong |
| author_sort | Juyong Park |
| collection | DOAJ |
| description | Wafer chips are manufactured in the semiconductor industry through various process technologies. Photolithography is one of these processes, aligning the wafer and scanning the circuit pattern on the wafer on which the photoresist film is formed by irradiating light onto the circuit pattern drawn on the mask. As semiconductor technology is highly integrated, alignment is becoming increasingly difficult due to problems such as reduction of alignment margin, transmittance due to level stacking structure, and an increase in wafer diameter in the photolithography process. Various methods and research to reduce the misalignment problem that is directly related to the yield of production are constantly being conducted. In this paper, we use machine vision for exposure equipment to improve the image resolution quality of marks for accurate alignment. To improve image resolution quality, we propose an improved Multi-Scale Residual Network (MSRN) that combines Attention Mechanism using a Multi-Scale Residual Attention Block to improve image resolution quality. Our proposed method can extract enhanced features using two different bypass networks and attention blocks with different scale convolution filters. Experiments were used to verify this method, and the performance was improved compared with previous research. |
| first_indexed | 2024-03-09T20:46:17Z |
| format | Article |
| id | doaj.art-076d1ee5b77b43aaabbd03463c352135 |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-09T20:46:17Z |
| publishDate | 2022-03-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-076d1ee5b77b43aaabbd03463c3521352023-11-23T22:45:18ZengMDPI AGApplied Sciences2076-34172022-03-01125272110.3390/app12052721Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in PhotolithographyJuyong Park0Jongpil Jeong1Department of Smart Factory Convergence, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon 16419, KoreaDepartment of Smart Factory Convergence, Sungkyunkwan University, 2066 Seobu-ro, Jangan-gu, Suwon 16419, KoreaWafer chips are manufactured in the semiconductor industry through various process technologies. Photolithography is one of these processes, aligning the wafer and scanning the circuit pattern on the wafer on which the photoresist film is formed by irradiating light onto the circuit pattern drawn on the mask. As semiconductor technology is highly integrated, alignment is becoming increasingly difficult due to problems such as reduction of alignment margin, transmittance due to level stacking structure, and an increase in wafer diameter in the photolithography process. Various methods and research to reduce the misalignment problem that is directly related to the yield of production are constantly being conducted. In this paper, we use machine vision for exposure equipment to improve the image resolution quality of marks for accurate alignment. To improve image resolution quality, we propose an improved Multi-Scale Residual Network (MSRN) that combines Attention Mechanism using a Multi-Scale Residual Attention Block to improve image resolution quality. Our proposed method can extract enhanced features using two different bypass networks and attention blocks with different scale convolution filters. Experiments were used to verify this method, and the performance was improved compared with previous research.https://www.mdpi.com/2076-3417/12/5/2721MSRNattention mechanismcomputer visionsuper-resolutionphotolithography |
| spellingShingle | Juyong Park Jongpil Jeong Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography Applied Sciences MSRN attention mechanism computer vision super-resolution photolithography |
| title | Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography |
| title_full | Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography |
| title_fullStr | Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography |
| title_full_unstemmed | Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography |
| title_short | Improved MSRN-Based Attention Block for Mask Alignment Mark Detection in Photolithography |
| title_sort | improved msrn based attention block for mask alignment mark detection in photolithography |
| topic | MSRN attention mechanism computer vision super-resolution photolithography |
| url | https://www.mdpi.com/2076-3417/12/5/2721 |
| work_keys_str_mv | AT juyongpark improvedmsrnbasedattentionblockformaskalignmentmarkdetectioninphotolithography AT jongpiljeong improvedmsrnbasedattentionblockformaskalignmentmarkdetectioninphotolithography |