Optical waveguide structures can make the state-of-the-art micro- and nanofabricated devices faster and less energy consuming. However, on-chip optical components must be placed at relatively large distances from each other, on the order of the wavelength $\lambda$, to eliminate the crosstalk between them. This makes on-chip optical structures too large to compete with their electronic counterparts. In this work, we explore the possibility of suppressing the crosstalk between closely spaced dielectric waveguides by making use of higher-order modes. We show that the crosstalk between two waveguides can be essentially eliminated, even if the waveguides are separated by a distance of $\lambda /10$ only. We also study arrays of more than two waveguides and find the conditions for the efficient crosstalk reduction in them. Our results can lead to further miniaturization of photonic integrated circuits.

Distributing entanglement through standard telecommunication fiber is particularly important for quantum key distribution protocols with low vulnerability over metropolitan distances. However, entanglement distribution over long distance through optical fiber so far could only be accomplished with moderate photon pair rates. In this work, we present entanglement distribution over 50 km of standard telecommunication fiber with a pair rate of more than 10 000 ${\mathrm{s}}^{-1}$ using a bright nondegenerate photon pair source. Signal and idler wavelengths of this source are optimized for low dispersion in optical fiber and high efficiency for single-photon avalanche diode detectors, respectively. The resulting modest hardware requirement and high rate of detected entangled photon pairs could significantly enhance practical entanglement-based quantum key distribution in existing metropolitan fiber networks.

In the past few years, the emerging photon-magnon dissipative coupling offers us an effective way to design and exploit losses to fulfill quantum tasks in hybrid cavity magnonics systems. In this paper, we theoretically investigate the dissipative-coupling-induced transparency (DCIT) and the enhancement of the high-order sidebands in a dissipative photon-magnon coupling system with Kerr nonlinearity. We find that the DCIT can be observed with both a weak and strong driving field. Due to the existence of the Kerr nonlinearity and introduction of the strong driving field of magnon, the sidebands can be enhanced highly by choosing the optimized parameters. In order to find the optimized parameters to enhance sidebands, we utilize the eigenvalues of the effective Hamiltonian to find the critical point of the stability region, near which sidebands can be enhanced in maximal degree.

One of the major challenges in the design of nonreciprocal devices is the application of magnetic biasing. This problem has motivated a flurry of work for the development of magnetless nonreciprocal devices, with spatiotemporal modulation being one of the most promising approaches to this end. The most fundamental nonreciprocal component is the gyrator, which can be the building block of other nonreciprocal components, like ciculators and isolators. Spatiotemporally modulated (STM) gyrators are typically based on a pair of frequency converters with different modulation phases, connected through a linear time-invariant network. Such gyrators are usually designed to operate with a single modulation sideband. In this paper we argue that this approach is suboptimal and show how adding just one more sideband can increase the bandwidth by $40\mathrm{\%}$ with a lower modulation frequency and the same modulation amplitude. We present an implementation of the proposed approach based on a double-balanced circuit with Wheatstone bridges of varactors, which provides large isolation between the input signal and the sidebands. We provide a theoretical and an experimental analysis of the circuit with excellent agreement to each other. In addition to introducing a topology of STM gyrators with improved bandwidth compared with existing designs, our results highlight a potential connection between the bandwidth of STM devices and their harmonic order, which might also find applications in other STM systems, like isolators, circulators, and photonic topological insulators.

Achromatic lenses are crucial to light focusing and imaging. However, conventional achromatic lenses are restricted by the Abbe diffraction limit. Although superresolution optical lenses have been demonstrated to overcome the limit, the correction of chromatic aberration in continuous broadband for superresolution lenses still remains challenging. To address this issue, we propose a generalized strategy of constructing a frequency-dependent phase profile of linear dispersion for achromatic superresolution focusing in a continuous broad bandwidth. Specifically, based on an arbitrary wavelength-independent discrete phase profile of a monochromatic superresolution lens optimized at a given single optical frequency, a frequency-dependent phase profile of an achromatic superresolution lens can be constructed as the sum of the given discrete phase profile and a linear dispersion phase term. Based on the proposed approach, broadband achromatic superresolution metalenses are designed and demonstrated in the terahertz regime. The proposed method can be applied to visible and infrared optical spectra, which could have promising potential in various superresolution-related applications.

The concept of photonic frequency-momentum ($\omega$-$q$) dispersion has been extensively studied in artificial dielectric structures such as photonic crystals and metamaterials. However, the $\omega$-$q$ dispersion of electrodynamic waves hosted in natural materials at the atomistic level is far less explored. Here, we develop a Maxwell Hamiltonian theory of matter combined with the quantum theory of atomistic polarization to obtain the electrodynamic dispersion of natural materials interacting with the photon field. We apply this theory to silicon and discover the existence of anomalous atomistic waves. These waves occur in the spectral region where propagating waves are conventionally forbidden in a macroscopic theory. Our findings demonstrate that natural media can host a variety of yet to be discovered waves with subnanometer effective wavelengths in the picophotonics regime.

Nonlinear Cherenkov radiation, an intriguing noncollinear optical parametric process with versatile longitudinal phase matching, is commonly realized in engineered nonlinear photonic structures. Here, we report on the observation of ultrabroadband rainbow Cherenkov second-harmonic generation (UBR CSHG) in a single periodically poled lithium niobate (PPLN) nonlinear crystal driven by a near-infrared high-peak-power ultrashort femtosecond pump laser. Multiple colored nonlinear radiation bands along the direction of Cherenkov conical angles are simultaneously exhibited, leading to a beautifully colorful rainbow picture never reported before. We systematically analyze the multicolor spectral characteristics of the UBR CSHG in the whole conical direction and explicitly clarify their phase-matching conditions. We show that this UBR CSHG interaction is subject to remarkable spectral broadening owing to both the third-order nonlinearity effect induced by the high-peak-power ultrashort pump laser and the reciprocal lattice vector of the PPLN crystal, which eventually enables the generation of three bright dominating colors—red, orange-yellow, and green bands—and their automatic dispersion in space. The experimental results suggest rich physics of nonlinear optical interactions and indicate the possibility to engineer broadband nonlinear frequency conversion for applications in compact colored laser sources and others.

We propose a free-space, inverse design of a nanostructure’s effective mode-matching fields via a backward propagation of tightly focused vector beams to the pupil plane of an aplanatic system of high numerical aperture. First, we study the nanostructure’s eigenmodes without considering any excitation fields and then extract the modal near fields in the focal plane. Each modal field is then taken as the desired focal field, the band-limited waves of which are backward propagated to the pupil plane via a reversal of the Richards-Wolf vector diffraction formula. The pupil fields can be designed to be genuinely paraxial by associating the longitudinal electric or magnetic field component with the radial one on the reference sphere. The inversely designed pupil field in turn is propagated forwardly into the focal region to generate the designed focal field, whose distribution over the nanostructure’s surface is used to evaluate the overlap between the designed focal field and the modal fields, i.e., the modal expansion coefficients. Studies for a silicon nanodisk monomer, dimer, and tetramer demonstrate the ability of our inverse approach to design the necessary tightly focused vector field that can effectively and exclusively match a certain eigenmode of interest. Compared with the forward beam-shaping method, the inverse design approach tends to yield quantitatively more precise mode-matching field profiles. This work can have a significant impact on optical applications that rely on controllable and tunable mode excitation and light scattering.

A quantum state transformation can be generally approximated by single- and two-qubit gates. This, however, does not hold with noisy intermediate-scale quantum technologies due to the errors appearing in the gate operations, where errors of two-qubit gates such as controlled-not and swap operations are dominated. In this work we present a cost efficient single-copy certification for a realization of a two-qubit gate in the presence of depolarization noise, where we aim to identify if the realization is noise free, or not. We show that entangled resources such as entangled states and a joint measurement are not necessary for the purpose, i.e., a noise-free two-qubit gate is not needed to certify an implementation of a two-qubit gate. We present a proof-of-principle demonstration with photonic qubits.

We theoretically demonstrate generation of the Stokes temporal localized structure (TLS) by exploiting its Raman interaction in space and time within an optical potential well shared with another dark pulse. Excitation of the Stokes TLSs is feasible in both the normal and anomalous group-velocity dispersion regime, with a single modulated pump source. Stimulated Raman scattering constitutes influences on the stability and physical feature of the generated Stokes TLS (STLS). Particularly, breathing STLSs are also observed due to the periodic energy transfer between the primary and the Stokes fields. These findings could deepen the understanding of complex nonlinear dynamics in resonators and facilitate the excitation of different types of TLSs in potential platforms.

The majority of sources of coherent optical radiation rely on laser oscillators driven by population inversion. Despite their technological importance in communications, medicine, industry, and other fields, it remains a challenge to access the spectral range of 0.1–10 THz (the “terahertz gap”), a frequency band for applications ranging from spectroscopy to security and high-speed wireless communications. Here, we propose a way to produce coherent radiation spanning the THz gap by efficient second-harmonic generation in low-loss dielectric structures, starting from technologically mature electronic oscillators in the approximately 100 GHz range. To achieve this goal, we introduce hybrid THz-band dielectric cavity designs that combine (1) extreme field concentration in high-quality-factor resonators with (2) nonlinear materials enhanced by phonon resonances. We theoretically predict conversion efficiencies of $>{10}^3\mathrm{\%}/$W and the potential to bridge the THz gap with 1 W of input power. This approach enables efficient, cascaded parametric frequency converters, and light sources extensible into the mid-IR spectrum and beyond.

Magic angles, namely, relative rotations between two parallel graphene flakes that create bigger periodic superlattices, are recently found to inherit the bilayer structure with exotic properties totally unknown to a single isolated graphite sheet. An electromagnetic analog to these regimes is provided by modeling the two monolayers of the same two-dimensional medium as anisotropic surface conductivities, which, if properly tilted, give substantial enhancement for the fields across the formed cavity. Such results may apply the paradigm of twistronics to nanophotonic design involving stacked optical materials, for serving a wide range of applications: from electromagnetic sensing and signal processing to radiative heat transfer and light harvesting.

The study of helicity in the context of light-matter interactions is an increasing area of research. However, some fundamental aspects of the helicity content of light fields inside scatterers have been overlooked. In this work, we demonstrate that the helicity of light fields inside lossless spherical cavities cannot be either conserved or sign flipped as a result of a scattering event. The underlying reason is that the internal electric and magnetic Mie coefficients cannot simultaneously oscillate with equal amplitude and equal or opposite phase. Our analytical demonstration is fulfilled regardless of the refractive index, sphere size, and multipolar order. In addition, we show that the helicity of light fields inside lossy spheres can be conserved. This fact is in striking contrast to the behavior of the scattered field, whose helicity content cannot be conserved precisely when the sphere has losses. Finally, we show that the helicity content of internal fields can be flipped for materials with gain.

The convolution operation is widely used in signal and image processing and represents the most computationally intensive step in convolutional neural networks. We introduce a scheme to achieve arbitrary convolution kernels in the synthetic frequency dimension with a simple setup consisting of a ring resonator incorporating a phase and an amplitude modulator. This scheme can be used to perform multidimensional convolutions. We provide an analytic approach that determines the required modulation profile for any convolution kernel. Our work points to a direction of using optical computing to remove the computational bottleneck in traditional electronic circuits and may be useful in improving machine-learning hardware in artificial-intelligence applications.