Prof. Wolfgang Osten is in charge of the Institute of Applied Optics at the University of Stuttgart and an established expert for industrial measurement and inspection processes. In this interview, he talks about the importance of quality assurance in networked production worlds, the balancing act between nano resolution and high inspection speeds in semiconductor production, and about the Optical Metrology conference that he will be in charge of for the last time at the World of PHOTONICS Congress from June 25 to 29 in Munich, Germany.
Prof. Wolfgang Osten: This time, the Optical Metrology Symposium, which is organized by the SPIE, will once again have six sub-conferences. Among other things, it will deal with the use of optical measurement methods in the industrial environment, their modeling in the course of inspection processes at extreme resolutions, and their specific use for inspecting artworks and biological materials. The focus will also be on exciting areas such as machine vision and the practical use of distance sensors. What began as two sub-conferences in 1999 has now become a symposium with six sub-conferences. This reflects the much increased significance of optical metrology. Its importance will grow even more in the context of digital transformation and, especially, the Industry 4.0 initiative. Metrology is a key component of industrial value creation. Only what can be measured can be produced and optimized. Today, industry simply cannot do without optical sensors and especially not in a digital transformation—which we will be discussing at the new SPIE conference Digital Optical Technologies.
Osten: In Digital Optical Technologies, representatives from major IT companies such as Microsoft, Nvidia, and Google are involved in the program committee—and we have around 75 registrations for presentations. This is a respectable achievement for a conference that we’re staging for the first time. And it underscores the importance of the entire spectrum of these technologies. The aim is to put the topic of digitization and its significance for optics into concrete terms and bring it to the attention of the general public. The topic ranges from virtual, augmented, and mixed reality, to computational optics and digital optics for sensor and authentication systems. We will discuss this area from the technical basics of production to the use of such systems. We are using this wide span to approach digital transformation from many different perspectives—and to develop new ways of thinking. My hope is that the new conference will further advance the WORLD of PHOTONICS Congress and the Munich location.
Osten: Allow me to make an observation. We have been discussing the role of photonics in an overall economic context for at least tow decades. The discussion is now being summed up under the heading of Photonics 4.0, and is experiencing a huge response from the public. I actually see only two main changes: networking and the rapid increase in processor performance. For example, in multi-sensor systems, they allow an enormous volume and variety of data to be recorded, processed, and evaluated within a very short time. Everything else—imaging, data processing, digital control, design and production planning—are established processes. What is new is that we now connect not only data but also things. Industrial Internet describes this more in more concrete terms than Industry 4.0. And it expresses what has to happen in industrial inspection: It gets closer to or inside the process. With networked sensors for inline quality control and real-time data analysis, it will be possible to control the processes automatically and indicate imminent faults at an early stage. In networked production, sensors are the interface between the digital-virtual and the real world.
Osten: We are experiencing a change toward flexibilization of production. Individualized products require quality assurance to be just as flexible as the production processes that it monitors. This calls for more broad-based, more flexible inspection systems than we have today, improved man-machine interfaces, and new solutions in machine vision, which has been a hot topic for decades. But we don’t have to reinvent the wheel for this. The aim is to expand the agenda of Photonics 2020 in a targeted manner and update it in the context of digitization. I think it’s important to focus more strongly on the life sciences, and to set new emphases in relation to innovative materials, multi-sensor systems and data fusion. And we should also include the growing opportunities offered by visualization and flexible production, e.g. through additive manufacturing. And, quite clearly, metrology is vital in all of these areas.
Osten: I don't think that should be our goal. 100-percent checks would be too expensive and time consuming. I also can’t imagine that it would be a milestone for the semiconductor industry. Structure sizes today are ten nanometers and will be reduced to eight and six nanometers in the coming years. You cannot scan that completely on a 450 mm diameter wafer. Instead, it’s about much more common, reliable random checks in all phases of wafer production. A very high number of different optical sensors are already used in today’s plants. The measurement effort is enormous. The required precision is ensured only in the interplay with nanopositioning, whose accuracy is in the sub-nanometer range, just like the measurement systems for lithography optics. The smaller the structures, the greater the influence of line edge roughness and optical blurring caused by refraction, scattering, etc. On top of this, chip producers are changing from 2D to 3D structures, which make the whole thing even more complex. Metrology and positioning technology are then no longer in the nano but rather in the pico range.