From March 17 to 19, 2021, SEMICON China, a grand event for the semiconductor industry, was held at the Shanghai New International Expo Center, attracting almost all the top technology providers in the semiconductor field at home and abroad. Jushi Technology brought the first set of fully deep learning-driven semiconductor lead frame defect detection system at home and abroad - "Poly core 2000" to the N5 main exhibition hall, and launched a variety of products: including general chip 2D/3D detection and measurement - poly core 3000, wafer level defect analysis ADC - poly core 5000. At the exhibition, Jushi Technology's Juxin series products have received high attention and praise from the industry.
Dr. Zheng Jun, founder and CEO of Jushi Technology, said at semicon China that in the field of semiconductor defect analysis and visual inspection, Jushi Technology is clearly positioned, that is, all products are currently aimed at domestic blanks and industry gaps.
After the Spring Festival in 2021 to the present, the Juxin 2000 market has performed strongly, and the number of orders in the first quarter has exceeded dozens. With the recent pull of China's semiconductor manufacturing capacity, the market demand for Juxin 2000 has continued to increase. So far, the top four manufacturers in China's domestic production capacity have become customers of Juxin 2000. At this SEMICON exhibition, Juxin 2000 also realized customer on-site orders. The rapid enlargement of the customer list of Juxin 2000 fully verifies the AI innovation value of Juxin 2000.
Juxin 2000 is an AI detection system developed by Jushi Technology for the high-density etching process process of lead frames, and is also the world's first full-depth learning-driven automatic optical inspection (AOI) equipment for semiconductor etching lead frames, breaking through the pattern that has always been dominated by American, Japanese and Korean manufacturers, and achieving high-quality localization substitution. Compared with the United States, Japan and South Korea products, Juxin 2000 has achieved multi-point innovation, especially in the primary testing indicators, the detection accuracy exceeds that of foreign counterparts by 10 times; in addition, Juxin 2000 also has unique innovative functions such as defect classification, quantitative detection, and cross-product transfer learning that foreign equipment does not have.
In addition to the performance advantages, the operating system of Juxin 2000 is extremely easy to use, and at the same time, it has AI intelligent guidance, convenient human-computer interaction HMI, and multi-machine interconnection for AI capability sharing, laying the foundation for the automation and unmanned production line of advanced semiconductor manufacturing. At the same time, superimposed on powerful industrial software functions and machine learning data analysis capabilities, Juxin 2000 effectively realizes the digital closed loop of semiconductor manufacturing quality management.
The product innovation of Juxin 2000 is not only a domestic substitution in a simple sense, but also realizes a number of innovative functions that foreign equipment does not have, and has been verified by large-scale production, market and customers, which fully proves that Juxin 2000 is a technological innovation ahead of foreign imported equipment.
Polycore 2000 equipment system
In addition to the display of Juxin 2000 real machine equipment, the semiconductor defect detection and analysis products launched by Jushi Technology on SEMICON also include Polycore 3000 and Polycore 5000.
Juxin 3000 is a general semiconductor chip 2D/3D defect detection system based on deep learning and integrated AI technology, which supports the detection and packaging of chips in QFP, QFN, BGA and other packaging forms. Based on the MatrixSemi platform of Jushi Technology, through deep learning, Juxin 3000 realizes high-speed and accurate measurement of 2D and 3D feature parameters of chips that exceed optical resolution, realizing high detection and low manslaughter of full surface appearance defects of chips, which can help customers improve the reliability of high-end chip shipments.
Juxin 5000 is a wafer-level advanced package ADC detection device based on machine learning, which can be deployed separately on the one hand to achieve refined AI-driven wafer defect detection and insight analysis. On the other hand, it can be "old" on a large scale, and can be cascaded with AOI equipment in the field of semiconductor packaging, making up for the lack of capacity of mainstream AOI equipment in foreign countries, and effectively improving the defect insight ability and quality control of advanced packaging production lines.