Wilmert De Bosscher, Soleras Advanced Coatings, Deinze, Belgium
Digitalization has come a long way and adoption into industry is happening at an ever-increasing pace. The revolution in high volume and large area vacuum coating applications is happening today at many different levels across our industry. This presentation tries to capture the various activities of digital transformation in industrial deposition processes and how these may significantly impact our expectations and way of working in the near-future coating businesses.
In the first part, we want to highlight some generic approaches allowing us to make better, higher performing and flexible coating equipment. It all started about 30 years ago with 3D CAD software, extended with some FEA (Finite Element Analysis) modules for making the most efficient mechanical constructions. The past 20 years, many have contributed in modelling gas flow, plasma and coating processes enabling to predict and better understand deposition performance in layer thickness, morphology and composition. In the past decade, deposition source positioning and substrate movement have allowed to anticipate forming uniform coating on more complex shaped substrate geometries. Today, generative AI (Artificial Intelligence) may allow proposing a coater equipment concept (e.g., number and size of chambers, geometry of coat zones, gas and pumping needs, infrastructure boundaries, …) fulfilling the expectations of the envisioned product performance and throughput.
The second part will focus on data management. Although most recent coaters have plenty of possibilities for logging sensor data of process parameters in real time, often data mining and correlation analysis is insufficiently explored. In addition, many companies develop coater components (e.g., pumps, power supplies, magnetrons, in-situ metrology tools, …) for which powerful standalone datalogging tools and dashboards are being developed. Again, limited integration into overall coater data management systems may limit the extended capabilities of those advanced tools. Having an appropriate IT framework and connection data gateways may enable linking all setting parameters and sensor data to the performance of individual coated products. A new and flexible control and monitor interface is desired, allowing to see exactly what you need, when you need it and with user specific (e.g., operator, R&D scientist, coater manager, …) configuration tools.
The third and final part discusses the capabilities of incorporating generative artificial intelligence into coater and coating performance while using aggregated data from prior operations. By linking historical metrology data from the coated product with related sensor and process data, performance of the coated product may be predicted during the coating process. Furthermore, critical variables may be tuned (single or multivariate analysis) in-situ to sustain the required coating performance. However, this is not a straightforward approach and requires acquisition of and providing context to all relevant data variables. In addition, component performance and the potential need for preventive maintenance on the coating equipment may be monitored and allowing optimal functionality of the system.
A combination of these approaches will enable us to optimize complex coating functionalities, at a desired high throughput, at minimized costs and while providing reliably high yield product performance.