Volume 8, Issue 3

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 complex models and algorithms help to exploit the enormous parameter space in a highly efficient way. In this review, we summarize the recent advances on the emerging field where nanophotonics and machine learning blend. We provide an overview of different computational methods, with the focus on deep learning, for the nanophotonic inverse design. The implementation of deep neural networks with photonic platforms is also discussed. This review aims at sketching an illustration of the nanophotonic design with machine learning and giving a perspective on the future tasks.

Nonlinear optical processes are vital for fields including telecommunications, signal processing, data storage, spectroscopy, sensing and imaging. As an independent research area, nonlinear optics began with the invention of the laser, because practical sources of intense light needed to generate optical nonlinearities were not previously available. However, the high power requirements of many nonlinear optical systems limit their use, especially in portable or medical applications, and so there is a push to develop new materials and resonant structures capable of producing nonlinear optical phenomena with low-power light emitted by inexpensive and compact sources. Acoustic nonlinearities, especially giant acoustic nonlinear phenomena in gas bubbles and liquid droplets, are much stronger than their optical counterparts. Here, we suggest employing acoustic nonlinearities to generate new optical frequencies, thereby effectively reproducing nonlinear optical processes without the need for laser light. We critically survey the current literature dedicated to the interaction of light with nonlinear acoustic waves and highly nonlinear oscillations of gas bubbles and liquid droplets. We show that the conversion of acoustic nonlinearities into optical signals is possible with low-cost incoherent light sources such as light-emitting diodes, which would usher new classes of low-power photonic devices that are more affordable for remote communities and developing nations, or where there are demanding requirements on size, weight and power.

Monitoring drug release in vitro and in vivo is of paramount importance to accurately locate diseased tissues, avoid inappropriate drug dosage, and improve therapeutic efficiency. In this regard, it is promising to develop strategies for real-time monitoring of drug release inside targeted cells or even in living bodies. Thus far, many multi-functional drug delivery systems constructed by a variety of building blocks, such as organic molecules, polymeric nanoparticles, micelles, and inorganic nanoparticles, have been developed for drug release monitoring. Especially, with the advancements in imaging modalities relating to nanomaterials, there has been an increasing focus on the use of non-invasive imaging techniques for monitoring drug release and drug efficacy in recent years. In this review, we introduce the application of fluorescence imaging, magnetic resonance imaging (MRI), surface-enhanced Raman scattering (SERS), and multi-mode imaging in monitoring drug release, involving a variety of nanomaterials such as organic or inorganic nanoparticles as imaging agents; their design principles are also elaborated. Among these, a special emphasis is placed on fluorescence-based drug release monitoring strategies, followed by a brief overview of MRI, SERS, and multi-mode imaging-based strategies. In the end, the challenges and prospects of drug release monitoring are also discussed.

We propose an electrically tunable metasurface, which can achieve relatively large phase modulation in both reflection and transmission modes (dual-mode operation). By integration of an ultrathin layer of indium tin oxide (ITO) as an electro-optically tunable material into a semiconductor-insulator-semiconductor (SIS) unit cell, we report an approach for active tuning of all-dielectric metasurfaces. The proposed controllable dual-mode metasurface includes an array of silicon (Si) nanodisks connected together via Si nanobars. These are placed on top of alumina and ITO layers, followed by a Si slab and a silica substrate. The required optical resonances are separately excited by Si nanobars in reflection and Si nanodisks in transmission, enabling highly confined electromagnetic fields at the ITO-alumina interface. Modulation of charge carrier concentration and refractive index in the ITO accumulation layer by varying the applied bias voltage leads to 240° of phase agility at an operating wavelength of 1696 nm for the reflected transverse electric (TE)-polarized beam and 270° of phase shift at 1563 nm for the transmitted transverse magnetic (TM)-polarized light. Independent and isolated control of the reflection and transmission modes enables distinctly different functions to be achieved for each operation mode. A rigorous coupled electrical and optical model is employed to characterize the carrier distributions in ITO and Si under applied bias and to accurately assess the voltage-dependent effects of inhomogeneous carrier profiles on the optical behavior of a unit cell.

We prove that, at any point of any reciprocal structure, the local field intensity enhancement, which is averaged over the incident wave polarizations and the incidence directions, exactly coincides with the radiative part of the Purcell factor, which is averaged over the orientations of the emitter. This result is important for gaining a better understanding of the electromagnetic Raman gain in the surface-enhanced Raman scattering and for the further development of this technique.

Atomically thin 2D materials such as transition metal dichalcogenides (TMDs) provide a wide range of basic building blocks with unique properties, making them ideal for heterogeneous integration with a mature chip platform for advances in optical communication technology. The control and understanding of the precise value of the optical index of these materials, however, is challenging, as the standard metrology techniques such as the millimeter-large ellipsometry is often not usable due the small lateral 2D material flake dimension. Here, we demonstrate an approach of passive tunable coupling by integrating few layers of MoTe 2 onto a microring resonator connected to a waveguide bus. We find the TMD-to-ring circumference coverage length ratio required to precisely place the ring into a critical coupling condition to be about 10% as determined from the variation of spectral resonance visibility and loss as a function of TMD coverage. Using this TMD-ring heterostructure, we further demonstrate a semiempirical method to determine the index of a 2D material ( n MoTe2 of 4.36+0.011 i ) near telecommunication-relevant wavelength. The placement, control, and optical property understanding of 2D materials with integrated photonics pave the way for further studies of active 2D material-based optoelectronics and circuits.

Conventional optical fiber has excellent performance in guiding light, which has been widely employed for long-distance optical communication. Although the optical fiber is efficient for transmitting light, its functionality is limited by the dielectric properties of the core’s and cladding’s materials (e.g. Ge-doped-silica and silica glasses). The spot size of the transmitted light is diverging and restricted by the diffraction limit of the dielectric core, and the numerical aperture is determined by the refractive index of the fiber materials. However, the novel technology of metasurfaces is opening the door to a variety of optical fiber innovations. Here, we report an ultrathin optical metalens directly patterned on the facet of a photonic crystal optical fiber that enables light focusing in the telecommunication regime. In-fiber metalenses with focal lengths of 28 μm and 40 μm at a wavelength of 1550 nm are demonstrated with maximum enhanced optical intensity as large as 234%. The ultrathin optical fiber metalens may find novel applications in optical imaging, sensing, and fiber laser designs.

Metasurfaces are two-dimensional arrangements of antennas that control the propagation of electromagnetic waves with a subwavelength thickness and resolution. Previously, metasurfaces have been mostly used to obtain the function of a single optical element. Here, we demonstrate a plasmonic metasurface that represents the combination of a phase mask generating a double-helix point spread function (DH-PSF) and a metalens for imaging. DH-PSF has been widely studied in three-dimensional (3D) super-resolution imaging, biomedical imaging, and particle tracking, but the current DH-PSFs are inefficient, bulky, and difficult to integrate. The multielement metasurface, which we label as DH-metalens, enables a DH-PSF with transfer efficiency up to 70.3% and an ultrahigh level of optical system integration, three orders of magnitude smaller than those realized by conventional phase elements. Moreover, the demonstrated DH-metalens can work in broadband visible wavelengths and in multiple incident polarization states. Finally, we demonstrate the application of the DH-metalens in 3D imaging of point sources. These results pave ways for realizing integrated DH-PSFs, which have applications in 3D super-resolution microscopy, single particle tracking/imaging, and machine vision.

Using coupled electromagnetic and hydrodynamic calculations, we elucidate theoretically the topographic transition from a random metallic surface to a periodic sub-wavelength grating by ultrashort laser ablation. The origin of this transition lies in the successive selection of hybrid surface waves scattered by random nanoholes. Contrary to the common belief that surface plasmon polaritons play the dominant role in the process and define the grating periodicity, we show that both quasi-cylindrical and surface plasmon waves are involved, whereas the diversity in the resulting spacings λ /2– λ ( λ is the laser wavelength) is the manifestation of a broad frequency overlap of these waves, controlled by their relative phase shifts with respect to the plasmonic counterparts. The topography evolution imposes the dominant contribution to the surface sub-wavelength pattern by selecting the appropriate wave character from plasmonic modes to evanescent cylindrical waves. With the radiation dose, the grating periodicity exhibits a pronounced blue shift due to reinforced dipole–dipole coupling between the nanoholes and surface curvatures in the laser-processed area. This allows the creation of regular patterns with tunable periodicity.

Polarization measurement has been widely used in material characterization, medical diagnosis and remote sensing. However, existing commercial polarization analyzers are either bulky schemes or operate in non-real time. Recently, various polarization analyzers have been reported using metal metasurface structures, which require elaborate fabrication and additional detection devices. In this paper, a compact and fully integrated silicon polarization analyzer with a photonic crystal-like metastructure for polarization manipulation and four subsequent on-chip photodetectors for light-current conversion is proposed and demonstrated. The input polarization state can be retrieved instantly by calculating four output photocurrents. The proposed polarization analyzer is complementary metal oxide semiconductor-compatible, making it possible for mass production and easy integration with other silicon-based devices monolithically. Experimental verification is also performed for comparison with a commercial polarization analyzer, and deviations of the measured polarization angle are <±1.2%.

Recent advances in opto-thermophoretic tweezers open new avenues for low-power trapping and manipulation of nanoparticles with potential applications in colloidal assembly, nanomanufacturing, life sciences, and nanomedicine. However, to fully exploit the opto-thermophoretic tweezers for widespread applications, the enhancement of their versatility in nanoparticle manipulations is pivotal. For this purpose, we translate our newly developed opto-thermophoretic tweezers onto an optical fiber platform known as opto-thermophoretic fiber tweezers (OTFT). We have demonstrated the applications of OTFT as a nanoparticle concentrator, as a nanopipette for single particle delivery, and as a nanoprobe. The simple setup and functional versatility of OTFT would encourage its use in various fields such as additive manufacturing, single nanoparticle-cell interactions, and biosensing.

The nonlinear optical microscopies of coherent two-photon excited fluorescence and anti-Stokes Raman scattering are strongly enhanced by multiple surface plasmon resonances (MSPRs). The Au@Ag nanorods presented strong MSPRs peaks at 800 and 400 nm, and can enhance nonlinear optical microscopy at fundamental and double frequencies, respectively. A two-dimensional (2D) material of g-C 3 N 4 is employed to study the plasmon-enhanced nonlinear optical microscopy by the femtosecond laser. The electric analysis reveals that the MSPRs of the Au@Ag nanorod can significantly enhance the signals of two-photon excited fluorescence and anti-Stokes Raman scattering by up to the orders of 10 4 and 10 16 , respectively. The results demonstrate the great advantages of plasmon-enhanced nonlinear optical microscopy for the optical analysis on 2D materials, thus providing a new adventure for increasing the optical resolutions of nonlinear optical microscopy.

We theoretically and numerically demonstrate that the spontaneous parity-time (PT) symmetry breaking phase transition can be realized respectively by using two independent tuning ways in a tri-layered metamaterial that consists of periodic array of metal-semiconductor Schottky junctions. The existence conditions of PT symmetry and its phase transition are obtained by using a theoretical model based on the coupled mode theory. A hot-electron photodetection based on the same tri-layered metamaterial is proposed, which can directly show the spontaneous PT symmetry breaking phase transition in photocurrent and possesses dynamical tunability and switchability. This work extends the concept of PT symmetry into the hot-electron photodetection, enriches the functionality of the metamaterial and the hot-electron device, and has varieties of potential and important applications in optoelectronics, photodetection, photovoltaics, and photocatalytics.

Nano-patterned colloidal plasmonic metasurfaces are capable of manipulation of light at the subwavelength scale. However, achieving controllable lithography-free nano-patterning for colloidal metasurfaces still remains a major challenge, limiting their full potential in building advanced plasmonic devices. Here, we demonstrate plasmonic field guided patterning (PFGP) of ordered colloidal metallic nano-patterns using orthogonal laser standing evanescent wave (LSEW) fields. We achieved colloidal silver nano-patterns with a large area of 30 mm 2 in <10 min by using orthogonal LSEW fields with a non-focused ultralow fluence irradiation of 0.25 W cm −2 . The underlying mechanism of the formation of the nano-patterns is the light-induced polarization of the nanoparticles (NPs), which leads to a dipole-dipole interaction for stabilizing the nano-pattern formation, as confirmed by polarization-dependent surface-enhanced Raman spectroscopy. This optical field-directed self-assembly of NPs opens an avenue for designing and fabricating reconfigurable colloidal nano-patterned metasurfaces in large areas.