Although artificial intelligence stands at the center of the Fourth Industrial Revolution, there are few examples of how it can be applied in actual industrial settings. This is because the commercialization and field implementation of industrial AI are that difficult. However, news has emerged that SK Hynix has set out to improve yield by introducing Gauss Labs’ AI solution, Panoptes VM, into its notoriously complex semiconductor manufacturing process.

SK Hynix announced on the 10th that it has introduced AI solutions into its semiconductor processes to improve production operation efficiency and yield.

Last November, Gauss Labs, an industrial AI startup invested in by SK Hynix, launched the virtual measurement AI solution software product Panoptes VM (Virtual Metrology). SK Hynix introduced Panoptes VM to its mass production fabs starting in December.

Panoptes VM is a virtual measurement AI solution that predicts manufacturing process results using sensor data. The name Panoptes is derived from 'Panoptes,' the all-seeing giant with a hundred eyes in Greek mythology. In other words, it signifies monitoring everything that happens during the manufacturing process.

The thin film deposition process, in which Panoptes VM was first introduced, is a critical process for depositing thin films onto wafers, and the thickness and refractive index of the thin film are directly related to the quality of the semiconductor. However, the process of measuring this in fine thin films requires a significant amount of time and resources, making it very difficult to perform full measurement.

To address this, SK Hynix introduced Panoptes VM into its thin film deposition process. Subsequently, by integrating the data analyzed by Panoptes VM with Advanced Process Control (APC), the company achieved a 21.5% improvement in the average process variation and enhanced yield. Moving forward, SK Hynix and Gauss Labs are considering expanding horizontal deployment to other processes beyond thin film deposition.

A virtual measurement model utilizing AI technology on actual data generated in the manufacturing process demonstrates prediction accuracy similar to that of actual measurement equipment. Through such virtual measurement, manufacturers can achieve the same effect as 100% full inspection, and SK Hynix anticipates that the potential benefits of utilizing this virtual measurement data are limitless.

Kim Young-han, CEO of Gauss Labs, stated, “Gauss Labs is tackling and solving challenging problems that can deliver meaningful effects and significant value to manufacturing sites through the best AI technology,” adding, “Starting with Panoptes VM, we will continue to develop products that will drive innovation in the manufacturing industry.”

Regarding the adoption of this AI solution, Kim Young-sik, Vice President of Manufacturing & Technology at SK Hynix, stated, “SK Hynix is striving to realize a more intelligent Smart Factory through collaboration with Gauss Labs,” and expressed his vision, adding, “We will continue to secure a technological advantage by integrating AI technology throughout semiconductor development and production, and the introduction of Panoptes VM is just the beginning.”