Deep learning in the context of nano-photonics is mostly discussed in terms of its potential for inverse design of photonic devices or nanostructures. Many of the recent works on machine-learning inverse design are highly specific, and the drawbacks of the respective approaches are often not immediately clear. In this review we want therefore to provide a critical review on the capabilities of deep learning for inverse design and the progress which has been made so far. We classify the different deep learning-based inverse design approaches at a higher level as well as by the context of their respective applications and critically discuss their strengths and weaknesses. While a significant part of the community's attention lies on nano-photonic inverse design, deep learning has evolved as a tool for a large variety of applications. The second part of the review will focus therefore on machine learning research in nano-photonics "beyond inverse design". This spans from physics informed neural networks for tremendous acceleration of photonics simulations, over sparse data reconstruction, imaging and "knowledge discovery" to experimental applications.
Data-driven approaches have been proposed as effective strategies for the inverse design and optimization of photonic structures in recent years. In order to assist data-driven methods for the design of topology of photonic devices, we propose a topological encoding method that transforms photonic structures represented by binary images to a continuous sparse representation. This sparse representation can be utilized for dimensionality reduction and dataset generation, enabling effective analysis and optimization of photonic topologies with data-driven approaches. As a proof of principle, we l…
We propose an approach to engineer quartic metamaterials starting from the desired photonic states. We apply our method to the design of the high-k asymptotics of metamaterials, extreme non-reciprocity and complex bi-anisotropic media.
Photonic inverse design typically seeks designs parameterized by binary arrays, where the values of each element correspond to the presence or absence of material at a particular point in space. Gradient-based approaches to photonic inverse design often include thresholding of non-binary optimization arrays; when the thresholding is sufficiently sharp, binary designs are obtained. However, difficulty can arise due to vanishing gradients for sharp thresholds, which cause optimization to stall. Here, we present hypersphere optimization, a new method of carrying out binarized photonic optimizatio…
Diamond hosts optically active color centers with great promise in quantum computation, networking, and sensing. Realization of such applications is contingent upon the integration of color centers into photonic circuits. However, current diamond quantum optics experiments are restricted to single devices and few quantum emitters because fabrication constraints limit device functionalities, thus precluding color center integrated photonic circuits. In this work, we utilize inverse design methods to overcome constraints of cutting-edge diamond nanofabrication methods and fabricate compact and r…
Photonic resonators allowing to confine the electromagnetic field in ultra-small volumes and with long decay times are crucial to a number of applications requiring enhanced nonlinear effects. For applications to integrated photonic devices on chip, compactness and optimized in-plane transmission become relevant figures of merit as well. Here we optimize an encapsulated Si/SiO$_2$ photonic crystal nanobeam cavity at telecom wavelengths by means of a global optimization procedure, where only the first few holes surrounding the cavity are varied to decrease its radiative losses. This strategy al…
Three-dimensional complete photonic bandgap materials or photonic crystals block light propagation in all directions. The rod-connected diamond structure exhibits the largest photonic bandgap known to date and supports a complete bandgap for the lowest refractive index contrast ratio down to $n_{high}/n_{low} \sim 1.9$. We confirm this threshold by measuring a complete photonic bandgap in the infrared region in Sn--S--O $(n\sim1.9)$ and Ge--Sb--S--O $(n\sim2)$ inverse rod-connected diamond structures. The structures were fabricated using a low-temperature chemical vapor deposition process via …
We present design guidelines for realizing side-coupled photonic crystal nanobeam (PCN) cavities with efficient coupling to low order resonances through adjusting the overlap of the PCN cavity mode and a feeding bus waveguide in both physical and k-space. We show that optimal side-coupled configurations function at a non-zero k-vector offset between the bus waveguide and PCN cavity modes. The straightforward design of side-coupled PCNs with high contrast resonances opens the door to their practical implementation in multiplexed, on-chip photonic devices.
Transport inhibition via Anderson localization is ubiquitous in disordered periodic lattices. However, in crystals displaying only flat bands disorder can lift macroscopic band flattening, removing geometric localization and enabling transport in certain conditions. Such a striking phenomenon, dubbed inverse Anderson transition and predicted for three-dimensional flat band systems, has thus far not been directly observed. Here we suggest a simple quasi one-dimensional photonic flat band system, namely an Aharonov-Bohm photonic cage, in which correlated binary disorder induces an inverse Anders…
The realization of ultra-compact passive silicon photonic devices is becoming more and more important for the future large-scale photonic integration as desired for many systems. Although some compact silicon photonic devices have been demonstrated by using inverse design, the device performance is still insufficient for real applications. Here, we propose and realize several representative ultra-compact advanced passive silicon photonic devices with decent performances by introducing subwavelength-grating (SWG) structures developed by our high-efficiency inverse design method. These devices a…
Silicon photonics is becoming a leading technology in photonics, displacing traditional fiber optic transceivers in long-haul and intra-data-center links and enabling new applications such as solid-state LiDAR (Light Detection and Ranging) and optical machine learning. Further improving the density and performance of silicon photonics, however, has been challenging, due to the large size and limited performance of traditional semi-analytically designed components. Automated optimization of photonic devices using inverse design is a promising path forward but has until now faced difficulties in…
The ability of photonic crystal waveguides (PCWs) to confine and slow down light makes them an ideal component to enhance the performance of various photonic devices, such as optical modulators or sensors. However, the integration of PCWs in photonic applications poses design challenges, most notably, engineering the PCW mode dispersion and creating efficient coupling devices. Here, we solve these challenges with photonic inverse design, and experimentally demonstrate a slow-light PCW optical phased array (OPA) with a wide steering range. Even and odd mode PCWs are engineered for a group index…
On chip exciting electric modes in individual plasmonic nanostructures are realized widely; nevertheless, the excitation of their magnetic counterparts is seldom reported. Here, we propose a highly efficient on chip excitation approach of the magnetic dipole mode of an individual split ring resonator (SRR) by integrating it onto a photonic crystal nanocavity (PCNC). A high excitation efficiency of up to 58% is realized through the resonant coupling between the modes of the SRR and PCNC. A further fine adjustment of the excited magnetic dipole mode is demonstrated by tuning the relative positio…
Recent advances in the design and fabrication of on-chip optical microresonators has greatly expanded their applications in photonics, enabling metrology, communications, and on-chip lasers. Designs for these applications require fine control of dispersion, bandwidth and high optical quality factors. Co-engineering these figures of merit remains a significant technological challenge due to design strategies being largely limited to analytical tuning of cross-sectional geometry. Here, we show that photonic inverse-design facilitates and expands the functionality of on-chip microresonators; theo…
Photonic crystal cavities with bowtie defects that combine ultra-high Q and ultra-low mode volume are theoretically studied for low-power nanoscale optical trapping. By harnessing the localized heating of the water layer near the bowtie region, combined with an applied alternating current electric field, this system provides long-range electrohydrodynamic transport of particles with average velocities of 30 $\mathrm{μm/s}$ towards the bowtie region on demand by switching the input wavelength. Once transported to a given bowtie region, synergistic interaction of optical gradient and attractive …
Efficient light coupling into integrated photonic devices is of key importance to a wide variety of applications. "Inverse nanotapers" are widely used, in which the waveguide width is reduced to match an incident mode. Here, we demonstrate novel, "double inverse" tapers, in which we taper both the waveguide height, as well as the width. We demonstrate that in comparison to regular inverse tapers, the double inverse tapers have excellent polarization-independent coupling. In addition, the optimum coupling is achieved with much larger taper dimension, enabling the use of photolithography instead…
Nanophotonics has been an active research field over the past two decades, triggered by the rising interests in exploring new physics and technologies with light at the nanoscale. As the demands of performance and integration level keep increasing, the design and optimization of nanophotonic devices become computationally expensive and time-inefficient. Advanced computational methods and artificial intelligence, especially its subfield of machine learning, have led to revolutionary development in many applications, such as web searches, computer vision, and speech/image recognition. The comple…
Topological photonic crystal waveguides can create edge states that may be more robust against fabrication disorder, and can yield propagation modes below the light line. We present a fully three-dimensional method to modify state-of-the-art designs to achieve a significant bandwidth improvement for lossless propagation. Starting from an initial design with a normalized bandwidth of 7.5% (13.4 THz), the modification gives more than 100% bandwidth improvement to 16.2% (28.0 THz). This new design is obtained using automatic differentiation enabled inverse design and a guided mode expansion techn…
We report on the design and the fabrication of 1D photonic crystal (PhC) nanobeam cavities with optical resonances around 800 nm, compatible with Rubidium, Cesium or Argon atomic transitions. The cavities are made of Indium Gallium Phosphide (InGaP) material, a III-V semi-conductor compound which has a large index of refraction ($n \simeq 3.3$) favoring strong optical confinement and small mode volumes. Nanobeam cavities with inline and side coupling have been designed and fabricated, and quality factors up to $2\times 10^4$ have been measured.
We provide a compact 200 line MATLAB code demonstrating how topology optimization (TopOpt) as an inverse design tool may be used in photonics, targeting the design of two-dimensional dielectric metalenses and a metallic reflector as examples. The physics model is solved using the finite element method, and the code utilizes MATLAB's fmincon algorithm to solve the optimization problem. In addition to presenting the code itself, we briefly discuss a number of extensions and provide the code required to implement some of these. Finally, we demonstrate the superiority of using a gradient-based met…
We present a scheme for on-chip optical transduction of strain and displacement of Graphene-based Nano-Electro-Mechanical Systems (NEMS). A detailed numerical study on the feasibility of three silicon-photonic integrated circuit configurations is presented: Mach-Zehnder Interferometer(MZI), micro-ring resonator and ring-loaded MZI. An index-sensing based technique using a Mach-Zehnder Interferometer loaded with a ring resonator with a moderate Q-factor of 2400 can yield a sensitivity of 28 fm/sqrt(Hz), and 6.5E-6 %/sqrt(Hz) for displacement and strain respectively. Though any phase sensitive i…
