Product Line Head – Semiverse™ Solutions Lam Research Fremont, CA, United States
Technology, and semiconductor technology in particular, has been advancing at a rapid and accelerating pace since the invention of the transistor over 75 years ago. We’ve now reached the point that smart manufacturing equipment, powered by advanced chip technology and artificial intelligence, is producing the next generation of computer chips. This has created a virtuous cycle of innovation, where computer chips and machine learning are helping to produce new, more powerful chips, but even faster and better than before.
In this talk, we will examine the use of digital twins and smart manufacturing equipment in the semiconductor industry, and how they are improving manufacturing productivity. We will review different types of digital twins used in semiconductor development, and discuss issues related to data sharing and the level of abstraction required for each of these different digital twins. We will also address the concepts of intelligence and smart equipment, and will provide specific examples of how smart fab tools enabled by machine learning (equipment digital twins) are being used to optimize high volume semiconductor manufacturing. In addition, we will discuss the concept of wafer process digital twins (virtual chip fabrication), and how wafer process twins are currently being deployed to accelerate time to yield. Machine learning is being used in many of these technology domains and is driving real-time feed-forward and feedback optimization to deliver significant improvements in semiconductor manufacturing.
Next, we will review a case study in process recipe development for a high aspect ratio etch operation. The number of potential process recipe combinations for this operation is quite high, and narrowing the solution space is difficult. We will discuss how Lam Research created a virtual environment to replicate this etch operation, and how we optimized process recipe creation using the virtual environment. In Lam’s study, we discovered the limitations of machine learning in process recipe development and demonstrated the value of complementing human intelligence and experience with machine learning during process recipe optimization. We will also briefly review an ongoing initiative to develop a true “digital twin fab” and the challenges of integrating all of the elements of the virtual fab.
We will conclude the talk by reviewing the general challenges of using digital twins in semiconductor manufacturing and propose solutions to meet those challenges.