Sculpt Custom Light For New Technologies

On a 10-cent coin-sized structure, researchers at INRS and the University of Sussex have created an optical device that, using artificial intelligence, controls the properties of light.

At the heart of biomedical imaging systems and a host of technologies, there are lasers that must produce a light with specific properties depending on the application. To achieve this, optical systems can shape and control the characteristics of light lasers, with certain limitations for the user. Today, researchers are able to miniaturize optical tools and group them together on what is called an optical chip. Holding on the tip of a finger, the optical chip is a real obstacle course for the light that passes through, changing the way its properties. The development of a chip offers, as unveiled by Professor Roberto Morandotti and an international research team in an article published inNature Communications , new scientific and technological opportunities.

Whether it is the duration of the light pulse produced by the laser, its shape, its color or other advanced parameters, we seek to increase ever more control of lasers: every feature on which we can acting opens a new range of application possibilities. However, up to now, this control was limited to certain characteristics of the pulses and required expensive devices, bulky or difficult to evolve.

To overcome these limitations, many proposals have been considered theoretically, as is the case with the use of several combined light pulses. Unfortunately, the use of the combination of light pulses is a major operational obstacle: the number of possible combinations is simply too high and complex to be treated by a traditional numerical or experimental approach.

However, this avenue of research had too much potential to abandon it. Taking advantage of the great versatility, stability and unique properties of miniature optical systems - the integrated optical chips - developed at INRS by Professor Morandotti and his team, Benjamin Wetzel of the University of Sussex has developed an approach to to divide and recombine laser pulses and thus to control in an unprecedented way their individual characteristics. As the main author of the article, he thus obtains a formidable range of possibilities: "From this conceptually simple approach, we have the possibility of expanding exponentially all of the possible combinations of the parameters of the system that we control. . By modifying only a few variables in this chip,36 configurations of pulse shapes to control our optical system, more than the estimated number of stars in the universe, "Wetzel illustrates.

These astronomical numbers demonstrate the obstacle in terms of computing power that this approach requires. To overcome this, the international team has ogled on the side of artificial intelligence. An automatic learning algorithm has been used to determine the best combinations of parameters to obtain precisely the desired type of light.

For the purposes of the study, as proof of concept, the light was manipulated to obtain supercontinuum. They are enlarged light spectra obtained by intense interactions between light and matter, now used in a large number of technological applications. As the results published in Nature Communicationsdemonstrate , the integrated optical chip, coupled with the use of the learning algorithm, produces precisely an optimal pulse pattern to provide researchers with the complex physical dynamics sought.

These exciting results will have an impact in many areas, both in basic and applied research, since many of the commonly used optical systems depend on the same phenomena as those deployed for supercontinuum generation. In addition, the proposed system is inexpensive, very compact and can evolve into more complex systems.

The work of the international research team could lead to the development of other intelligent optical systems using self-optimization techniques. These may include the control of optical frequency combs (2005 Nobel Prize) for metrology applications, self-tuned lasers, light pulse processing and amplification (Nobel Prize 2018) and the execution of more fundamental approaches to intelligent learning such as systems based on photonic neural networks.

About This Study
Benjamin Wetzel, Michael Kues, Piotr Roztocki, Christian Reimer, Pierre-Luc Godin, Maxwell Rowley, Brent E. Little, Sai T. Chu, Evgeny A. Viktorov, David Moss, Alessia Pasquazi, Marco Peccianti and Roberto Morandotti, " Customizing supercontinuum generation via on-chip adaptive temporal pulse-splitting , " Nature Communications 9 , 4884 (2018).

The published work is the result of a collaboration between the National Institute for Scientific Research (INRS) and the University of Sussex (United Kingdom). The experimental work was carried out at INRS in Professor Morandotti's laboratory as part of Marie Curie international funding. The international team consists of researchers from INRS (Canada), the University of Sussex (United Kingdom), the Chinese Academy of Sciences (China), the City University of Hong Kong (China), ITMO University (Russia) and Swinburne University of Technology (Australia).

The research team received financial support from the Natural Sciences and Engineering Research Council of Canada , the Quebec Ministry of the Economy, Science and Innovation , the Canada Research Chairs , the Conseil de la Recherche Research Council, the European Research Council, the European Union, the Physical Sciences and Engineering Research Council (UK), the Government of the Russian Federation and the Sichuan 1000 Talent Program (China).