Metamaterials composed of metallic antennae arrays are used as they possess extraordinary optical transmission (EOT) in the terahertz (THz) region, whereby a giant forward light propagation can be created using constructive interference of tunneling surface plasmonic waves. However, numerous applications of THz meta-devices demand an active manipulation of the THz beam in free space. Although some studies have been carried out to control the EOT for the THz region, few of these are based upon electrical modulation of the EOT phenomenon, and novel strategies are required for actively and dynamically reconfigurable EOT meta-devices. In this work, we experimentally present that the EOT resonance can be coupled to optically reconfigurable chalcogenide metamaterials which offers a reversible all-optical control of the THz light. A modulation efficiency of 88% in transmission at 0.85 THz is experimentally observed using the EOT metamaterials, which is composed of a gold (Au) circular aperture array sitting on a non-volatile chalcogenide phase change material (Ge2Sb2Te5) film. This comes up with a robust and ultrafast reconfigurable EOT over 20 times of switching, excited by a nanosecond pulsed laser. The measured data have a good agreement with finite-element-method numerical simulation. This work promises THz modulators with significant on/off ratios and fast speeds.Metamaterials composed of metallic antennae arrays are used as they possess extraordinary optical transmission (EOT) in the terahertz (THz) region, whereby a giant forward light propagation can be created using constructive interference of tunneling surface plasmonic waves. However, numerous applications of THz meta-devices demand an active manipulation of the THz beam in free space. Although some studies have been carried out to control the EOT for the THz region, few of these are based upon electrical modulation of the EOT phenomenon, and novel strategies are required for actively and dynamically reconfigurable EOT meta-devices. In this work, we experimentally present that the EOT resonance can be coupled to optically reconfigurable chalcogenide metamaterials which offers a reversible all-optical control of the THz light. A modulation efficiency of 88% in transmission at 0.85 THz is experimentally observed using the EOT metamaterials, which is composed of a gold (Au) circular aperture array sitting on a non-volatile chalcogenide phase change material (Ge2Sb2Te5) film. This comes up with a robust and ultrafast reconfigurable EOT over 20 times of switching, excited by a nanosecond pulsed laser. The measured data have a good agreement with finite-element-method numerical simulation. This work promises THz modulators with significant on/off ratios and fast speeds.

Photonic inverse design concerns the problem of finding photonic structures with target optical properties. However, traditional methods based on optimization algorithms are time-consuming and computationally expensive. Recently, deep learning-based approaches have been developed to tackle the problem of inverse design efficiently. Although most of these neural network models have demonstrated high accuracy in different inverse design problems, no previous study has examined the potential effects under given constraints in nanomanufacturing. Additionally, the relative strength of different deep learning-based inverse design approaches has not been fully investigated. Here, we benchmark three commonly used deep learning models in inverse design: Tandem networks, Variational Auto-Encoders, and Generative Adversarial Networks. We provide detailed comparisons in terms of their accuracy, diversity, and robustness. We find that tandem networks and Variational Auto-Encoders give the best accuracy, while Generative Adversarial Networks lead to the most diverse predictions. Our findings could serve as a guideline for researchers to select the model that can best suit their design criteria and fabrication considerations. In addition, our code and data are publicly available, which could be used for future inverse design model development and benchmarking.Photonic inverse design concerns the problem of finding photonic structures with target optical properties. However, traditional methods based on optimization algorithms are time-consuming and computationally expensive. Recently, deep learning-based approaches have been developed to tackle the problem of inverse design efficiently. Although most of these neural network models have demonstrated high accuracy in different inverse design problems, no previous study has examined the potential effects under given constraints in nanomanufacturing. Additionally, the relative strength of different deep learning-based inverse design approaches has not been fully investigated. Here, we benchmark three commonly used deep learning models in inverse design: Tandem networks, Variational Auto-Encoders, and Generative Adversarial Networks. We provide detailed comparisons in terms of their accuracy, diversity, and robustness. We find that tandem networks and Variational Auto-Encoders give the best accuracy, while Generative Adversarial Networks lead to the most diverse predictions. Our findings could serve as a guideline for researchers to select the model that can best suit their design criteria and fabrication considerations. In addition, our code and data are publicly available, which could be used for future inverse design model development and benchmarking.

In this review we consider the development of optical near-field imaging and nanostructuring by means of laser ablation since its early stages around the turn of the century. The interaction of short, intense laser pulses with nanoparticles on a surface leads to laterally tightly confined, strongly enhanced electromagnetic fields below and around the nano-objects, which can easily give rise to nanoablation. This effect can be exploited for structuring substrate surfaces on a length scale well below the diffraction limit, one to two orders smaller than the incident laser wavelength. We report on structure formation by the optical near field of both dielectric and metallic nano-objects, the latter allowing even stronger and more localized enhancement of the electromagnetic field due to the excitation of plasmon modes. Structuring with this method enables one to nanopattern large areas in a one-step parallel process with just one laser pulse irradiation, and in the course of time various improvements have been added to this technique, so that also more complex and even arbitrary structures can be produced by means of nanoablation. The near-field patterns generated on the surface can be read out with high resolution techniques like scanning electron microscopy and atomic force microscopy and provide thus a valuable tool—in conjunction with numerical calculations like finite difference time domain (FDTD) simulations—for a deeper understanding of the optical and plasmonic properties of nanostructures and their applications.In this review we consider the development of optical near-field imaging and nanostructuring by means of laser ablation since its early stages around the turn of the century. The interaction of short, intense laser pulses with nanoparticles on a surface leads to laterally tightly confined, strongly enhanced electromagnetic fields below and around the nano-objects, which can easily give rise to nanoablation. This effect can be exploited for structuring substrate surfaces on a length scale well below the diffraction limit, one to two orders smaller than the incident laser wavelength. We report on structure formation by the optical near field of both dielectric and metallic nano-objects, the latter allowing even stronger and more localized enhancement of the electromagnetic field due to the excitation of plasmon modes. Structuring with this method enables one to nanopattern large areas in a one-step parallel process with just one laser pulse irradiation, and in the course of time various improvements have been added to this technique, so that also more complex and even arbitrary structures can be produced by means of nanoablation. The near-field patterns generated on the surface can be read out with high resolution techniques like scanning electron microscopy and atomic force microscopy and provide thus a valuable tool—in conjunction with numerical calculations like finite difference time domain (FDTD) simulations—for a deeper understanding of the optical and plasmonic properties of nanostructures and their applications.