Automated defectoscopy of thin poly (methyl methacrylate) layers

0567284 - ÚPT 2023 RIV CZ eng J - Journal Article
Podstránský, Jáchym - Knápek, Alexandr
Automated defectoscopy of thin poly (methyl methacrylate) layers.
Electroscope. Roč. 2022, č. 1 (2022), č. článku 8. ISSN 1802-4564
Institutional support: RVO:68081731
Keywords : defects in resist * artificial intelligence * image processing * automatization
OECD category: Automation and control systems
Method of publishing: Open access
http://147.228.94.30/images/PDF/Rocnik2020/Cislo1_2022/r16c1c7.pdf

In the electron beam lithography process, one of the initial steps is to coat the substrate (i.e., the silicon wafer) with a thin layer of polymer resist. During the coating process, defects in the thin layer can occur, which can affect the exposure and therefore the functionality of the final nanostructure. By checking the quality of the deposited polymer layer prior to exposure, these defect sites can be avoided. This process can be done manually using a visible-light microscope, but it is a time-consuming process and subject to a possible human error. In the framework of this project, a fully automated device has been developed that can detect and identifies these using computer vision. It is a scanning device that, by combining three stepper motors and an optical camera, takes images of the desired area of the wafer and then analyses these with the help of artificial intelligence. The user is then provided with a document in which the size, position and type of each defect found is recorded.
Permanent Link: https://hdl.handle.net/11104/0338550