Benoit Cluzel

Machine learning (ML) has become an important tool in nanophotonics for predicting the optical response of plasmonic nanostructures based on parameters, such as geometry and material. Traditional modeling approaches often involve high computational costs and long simulation times, which hinder rapid optimization and design. This work investigates the plasmonic resonance behavior of cylindrical gold nanostructures and evaluates several ML regression models for direct and inverse design. Among the tested methods, the Extra Trees Regressor (ET) achieved the best performance in predicting scattering spectra, with MAE of 0.097, MSE of 0.052, RMSE of 0.224, and R ² of 0.999. For inverse design, which maps spectral inputs to physical parameters, the model demonstrated high accuracy and strong generalization, recovering nanocylinder radius with errors below 2.6%. These results highlight the potential of ML-based strategies to accelerate nanophotonic design, reduce dependence on high-cost simulations, and support efficient characterization workflows. Graphical abstract

Interferometry has long been used to measure the phase of light signals. Combined with a heterodyne detection scheme, it allows to simultaneously record amplitude and phase variations. In this work, these well‐known techniques are exploited to evaluate the nonlinear phase induced by the optical Kerr effect during the propagation of a laser pulse in a nonlinear medium. It is shown that the nonlinear index can easily be retrieved when the accumulated phase remains small, but counter‐intuitive results can be observed at higher powers. In particular, the combination of pulsed excitation and Kerr nonlinearity results in nonlinear variations of the amplitude of the interference pattern, whose shape depends on the pulse profile and the dispersion of the propagation medium.

With the rise of optical computing, phase‐change materials (PCM) are attractive candidates for implementing programmable optical functionalities in photonic circuits. Integrating them as tunable media in photonic waveguides appears to be a simple but effective way to program their optical transfer function in both amplitude and phase through near‐field interactions. This is exactly what is being addressed in this work by integrating Ge₂Sb₂Te₅ (GST) patches on top of silicon waveguides. A scalable fabrication process is described, and the optical properties and programming capabilities are reported obtained from heterodyne interferometry to access the complex transmission of waveguides on the fly. It is reported that both amorphization and crystallization energy thresholds are at least an order of magnitude lower than previous reports, and programming endurance is successfully tested up to 10⁸ cycles, exceeding all previous results obtained with GST or any other PCM for use in photonics. Finally, the morphological and chemical characterization of PCM against aging is presented, revealing a failure mechanism due to phase segregation and ope ning new avenues to further optimize device endurance. As such, these results will be valuable to the large community of scientists seeking to integrate phase‐change materials into advanced photonic architectures where reconfigurability is required.

Plasmonic nanostructures have been widely studied for their unique optical properties, which are useful in sensing, photonics, and energy. However, the efficient design of these structures, considering the complex relationship between geometry, material, and optical response, remains a challenge. In this study, we propose a machine learning-based approach to address the inverse design problem in nanostructures, using data generated by numerical simulations via the Finite Element Method (FEM). We used a dataset of over 140,000 entries to train the regression models CatBoost, Random Forest, and Extra Trees, capable of predicting physical parameters, such as the radius of the nanocylinder, based on the simulated optical response. The CatBoost model achieved the best performance, with a Mean Absolute Error below 0.3 nm on unseen data. In parallel, we applied a direct design approach to experimental data of metallic nanoparticles, focusing on the optical absorption prediction from particle size. In this case, Random Forest presented the best performance, with a lower risk of overfitting. The results indicate that machine learning models are promising tools for optimizing the design and characterization of plasmonic nanostructures, thus reducing the need for costly experimental techniques.

We have realized a graded flat lens with negative effective index operating at telecom wavelength for Silicon Photonics. We report on its design, fabrication, and experimental characterization. This photonic device behaves as a convergent lens; it has been designed so as to be fabricated on a Silicon on Insulator platform and to operate at 1.55 μm. It consists of a graded photonic crystal whose filling factor has been varied so as to fit the profile of index of the graded index lens. Since it operates in the second band of Photonic Crystals, the effective index is negative. It has been characterized in two steps: (i) first, the output beam has been scanned by a set of waveguides and a CCD camera. (ii) Second, the focusing has been investigated by Scanning Near-field Optical Microscopy. This characterization validates the design: the focal length is about 5 μm (≃ 3λ), while the width of the focal spot is 0.87 μm (0.53 λ). This compact photonic device is at the wavelength scale (thickness = 4.4λ). Our design, based on Graded Index Optics, may be adapted to design new devices perfectly suited for Photonics applications, such as a short taper or a mode converter.

The highly promising linear and nonlinear optical properties of innovative thin films of GeSe_1–xTe_x chalcogenide materials in the amorphous as‐deposited state, and after crystallization, are revealed here. These innovative alloys bridge the gap between two families of materials: chalcogenide glasses (GeSe) and phase‐change materials (GeTe). Their unique optical properties make them attractive candidates for reconfigurable and nonlinear photonic applications in the infrared. In this context, it is shown how, by varying the Te content of GeSe_1–xTe_x thin films, it is possible to tailor their linear and nonlinear optical properties to optimize them for a wide range of innovative applications.

Three centuries ago, Isaac Newton reported the experimental observation of the tunnelling of a light ray between two prisms separated by a small gap once one of them was shinned in total internal reflection. This article describes a modern revisit of this seminal Newton’s experiment, generally known as the Frustrated Total Internal Reflection. This experiment was created in the framework of a series of lectures about near-field optics and nanophotonics for Master students. During a 4h lab work session, the students are running the experiment to evidence and quantify the evanescent wave on top of a glass prism once illuminated above the critical angle.

Metafibers expand the functionalities of conventional optical fibers to unprecedented nanoscale light manipulations by integrating metasurfaces on the fiber tips, becoming an emerging light-coupling platform for both nanoscience and fiber optics communities. Mostly exploring the isolated bare fibers, current metafibers remain as proof-of-concept demonstrations due to a lack of standard interfaces with the universal fiber networks. Here, we develop new methodologies to fabricate well-defined plasmonic metasurfaces directly on the end facets of commercial single mode fiber jumpers using standard planar technologies and provide a first demonstration of their practical applications in the nonlinear optics regime. Featuring plug-play connections with fiber circuitry and arbitrary metasurfaces landscapes, the metafibers with tunable plasmonic resonances are implemented into fiber laser cavities, yielding all-fiber sub-picosecond (minimum 513 fs) soliton mode locked lasers at optical wavelengths of 1.5 micrometer and 2 micrometer, demonstrating their unusual polarimetric nonlinear transfer functions and superior saturation absorption responses. Novel insights into the physical mechanisms behind the saturable absorption of plasmonic metasurfaces are provided. The nanofabrication process flow is compatible with existing cleanroom technologies, offering metafibers an avenue to be a regular member of functionalized fiber components. The work paves the way towards next generation of ultrafast fiber lasers, optical frequency combs, optical neural networks and ultracompact "all-in-fibers" optical systems for sensing, imaging, communications, and many others.

Semiconductor nanostructures are used in many applications, from electronics to nanophotonics to cite just a few, and most of these applications are based on silicon or silicon‐related technologies. With the miniaturization of these nanoscale devices, heat dissipation becomes a limitation to their efficiency, robustness, and reliability. Hence, managing thermal effects at nanoscale is attracting a growing interest and a promising approach relies on the control of acoustic phonons that are responsible of heat transfers. However, these phonons in silicon have high frequencies in the terahertz range which makes them hard to control and detect, thus limiting basic investigation and practical implementations. In this work, employing coherent control of femtosecond laser pulses and inelastic scattering spectroscopy, it is reported on the coherent generation, the control, and the detection of acoustic phonons in silicon metasurfaces in which light–matter interactions are enhanced.