公开(公告)号：US20240387552A1

公开(公告)日：2024-11-21

申请号：US18643149

申请日：2024-04-23

Applicant: Samsung Electronics Co., Ltd.

Inventor： Kyenhee Lee ,  Kyungsoo Kim

IPC: H01L27/118 ,  H03K19/20

Abstract: A semiconductor device includes: an active pattern extending on a substrate in a first direction; first to fourth channel structures stacked, in order, on one region of the active pattern; first to fourth gate structure respectively crossing the first to fourth channel structures, and extending in a second direction; first to fourth source/drain patterns, respectively, connected to both ends of the first to fourth channel structures; a plurality of upper contact vias electrically connecting each of a plurality of upper wiring lines to at least one of the first to fourth source/drain patterns; a plurality of lower wiring lines disposed on a lower surface of the substrate; and a plurality of lower contact vias penetrating through the substrate and electrically connecting each of the plurality of lower wiring lines to at least one of the first to fourth source/drain patterns.

公开(公告)号：US12229944B2

公开(公告)日：2025-02-18

申请号：US17842206

申请日：2022-06-16

Applicant: Samsung Electronics Co., Ltd.

Inventor： Kyenhee Lee ,  Mincheol Kang ,  Sooryong Lee

IPC: G06T7/00 ,  G06N3/04 ,  G06N3/045 ,  G06T7/11

Abstract: The inventive concept provides a defect detection method of a semiconductor element, capable of promptly and accurately detecting a defect, and predicting a type of the defect with respect to various defects of the semiconductor element, and a semiconductor element manufacturing method including the defect detection method. The defect detection method is capable of promptly and accurately detecting the defect, and predicting the type of the defect with respect to various defects of the semiconductor element, by generating a first segmentation image and a second segmentation image; converting the first segmentation image and the second segmentation image into an image of a first color and a second color, respectively; generating a combination image; classifying the type of a defect; generating a defect detection model by using deep learning, and detecting a defect of the semiconductor element by using a defect detection process using the defect detection model.