February 20, 2018
“Production with light is the key to Industry 4.0”
Prof. Michael Schmidt is Managing Director of the Bavarian Laser Center (blz). He heads the chair for photonic technologies at the University of Erlangen, is also the current President of Wissenschaftliche Gesellschaft Lasertechnik e.V. (WLT) and has extensive contacts in the global photonics sector. In the interview, he talks about the outlook for Photonics 4.0 in the health sector and industry, the potential of additive manufacturing and the use of ultrashort pulse lasers in materials processing and biomedical applications.
Prof. Schmidt, what does Photonics 4.0 mean for you?
To answer this question, I have to digress and talk about Industry 4.0—the trend toward intelligent, self-configuring process chains in which connected machines recognize parts and guide them individually through the production process. Then there is status and process monitoring using big data methods. These visions are from the world of IT and are only possible to a limited extent with current forming or joining technology. In contrast, Photonics 4.0 is sufficiently flexible. You simply reprogram a laser and it then produces the next product. Production with light is the key to Industry 4.0. After all, light can be used not only to process images, but also to measure, detect and provide them. Measuring, sensor and imaging data are indispensable in Industry 4.0 in order to control processes and monitor machine states. And in medicine too, optical methods and imaging processes are helping to constantly improve diagnostic and therapy options. Photonics 4.0 is the intelligent combination of optical technologies and smart software.
In the future, will software-based connectivity between photonic technologies be more important than new developments in the hardware field?
No, definitely not. We need both. The manipulation of parts or tissues with light by no means exhausts all the possibilities of light. Ultrashort pulse lasers are a good example. They have been gaining ground since the middle of the 2000s because they allow materials to be processed cold. There will also be further developments, such as the use of light polarization to allow more precise laser material processing. Or the intelligent control of ultrashort pulse lasers to influence the sequence of wavelengths and hence the material processing process as a whole. I always say that in the field of laser technology we have so far primarily taken a sledgehammer approach. Only now are we gradually beginning to use a scalpel. However, intelligent connectivity between new and existing hardware is also becoming increasingly important—in order to enable data-based optical processes.
The number of medical imaging procedures is rising. Doctors today find it difficult to select the right procedure and interpret the images. Can photonics 4.0 solve this dilemma?
When it comes to optical procedures, we are getting better and better at combining geometric and functional imaging, which is a positive start. An OCT (optical coherence tomograph) provides very precise information on the geometric make-up of eye tissues. When combined with two-photon processes, or pump probe spectrometry in the future, it offers the opportunity to examine the function of these tissues at the same time, hence providing me with much more precise information regarding problems in a patient’s body. It would be ideal if we had general databases in which all images of all patients worldwide are stored anonymously. We could then use big data methods and deep learning approaches to learn how to distinguish pathological tissue changes from those that are benign—on a largely automated basis ...
... even anonymized image data is a no-go for data privacy reasons. Are changes in the law needed in order to provide the necessary database for learning systems?
We have to make do with alternatives. It might be enough to store specific extracts showing pathological features instead of entire images in such a database. However, we would have to know first exactly what we are looking for. In the future, we will possibly increasingly use more screening methods at once on patients in examinations. The advantage of this multimodal imaging would be that we would need to visit the doctor less often and there would be a database for each patient that can be used to compare future images and identify abnormalities with the aid of algorithms.
They offer huge potential, especially when it comes to additive procedures in the area of metals. The electron beam procedure is still quicker and the microstructures of components can be influenced more accurately. Laser procedures, however, are still in their infancy. To date, the scanner mirrors have been a weakness as they restrict the processing speeds to just a few hundred meters per second. Research into alternatives is in full swing. We have already come up with ways of deflecting the laser beam masslessly with light or forming it as a whole. The latter option requires new, very high-performance lasers which expose the entire powder bed at once with a suitably programmable mask. Components would still be printed layer for layer but using a high-energy “laser flash” per layer rather than “dancing laser beams”. We may also see combinations of various wavelengths. New options might arise in the area of high-performance direct diode or quantum cascade lasers. And naturally, there is also considerable room for improvement when it comes to process control.
Along with photonics 4.0 and 3D printing, ultrashort pulse lasers are also in vogue. What, in your opinion, are the most important areas of application for ultrashort pulse lasers?
The key advantages are the possibility of very precise cold processing and the targeted processing of transparent media, because I can generate a plasma very locally which absorbs energy as the electrons travel. This is useful in eye surgery and is already widespread, with femtosecond and, increasingly, UV femtosecond lasers being used. What is lacking for large-scale industrial applications are systems with a moderate power output in the kilowatt range in order to increase the power per pulse. To ensure that this does not lead to undesirable warming, we need new ideas for beam deflection or beam shaping as is the case in additive manufacturing. This would enable us to use ultrashort pulse lasers to process large areas at once cold and with megahertz rates. We already have suitable approaches and ideas, but it will still take a few more years to make sufficient progress in macro areas. To sum up, we need new ways of deflecting light effectively and very quickly in all areas—ideally using massless methods. This would be a real breakthrough and, as is always the case with light, would have an effect in both directions—not only for processing but also for imaging and sensor systems.