J. Foucher, S. Martinez, H. Ozdoba, J. Baderot, A. Hallal, M. Jacob, S. Garrais
POLLEN Metrology Inc.,
United States
Keywords: process control, deep learning, software, software development kit, metrology, inspection, yield
Summary:
The complexity of semiconductor advanced processes development and manufacturing requires very challenging timelines to shorter time to market cycles. Despite a great evolution of tool hardware for process, metrology, inspection, the semiconductor industry needs a full digital transformation in daily operations to help process, characterization and yield engineers to work with more efficiency to better understand their advanced processes and optimize faster the fine tuning of process control from Lab to Fab. As long as new devices are introduced, the available data for process control are highly heterogeneous, coming from different imaging and non-imaging techniques for both metrology and inspection (SEM, TEM, Optical, Acoustic, AFM…). Consequently, more complex analyses take more time and are human dependent for fine tuning. The impact is huge on R&D costs raising and on production yield long ramp-up. Though, flexibility, automation and robustness are key topics. The industry is facing several key challenges: firstly, versatility of metrology and inspection capabilities has a major impact to handle a large variety of heterogeneous data. Secondly, specifications of measurements are evolving very quickly. A flexible solution where users can customize their own software to fit their specific environment is required and leverages internal data science work that is becoming more strategic. It enables users to be more efficient while maintaining consistency and quality of the extracted data. This final aspect also preserves the confidentiality of the data as the whole process can be performed locally. This is the real digital transformation requested for daily process development operations enhancement. To answer to these challenges, we are proposing a Collaborative and Holistic AI driven process control software platform tackling IC process control complexity. For the versatility of analysis, we powered the system with a generic AI pipeline that can recognize objects taught by the user with few examples and a system of geometric constraints to define a custom measurement recipe to characterize accurately the object. We validate the approach on a wide range of objects from different applications. For instance, the semiconductor industry includes logic and memory from early stage to late integration. To be compatible with IC Fabs specific needs and digital transformation, the software technology is collaborative and can be turned into a SDK (Software Development Kit) that allows any type of customization. It allows any data scientists to create their own plugin to easily tune the platform for specific environments such as R&D, production, built-in software for manufacturing equipment or microscope. It also gives an answer to scalability needs. Thus, there is a full synergy between data scientists and process engineers to maximize daily operations. We will demonstrate the capabilities of the software for both metrology and inspection applications on several semiconductor use cases showing the impact of state of the art deep learning library in combination with a collaborative and modular software architecture to accelerate time to market of IC Manufacturing.