Three-dimensional (3D) nonlinear photonic crystals have received intensive interest as an ideal platform to study nonlinear wave interactions and explore their applications. Periodic fork-shaped gratings are extremely important in this context because they are capable of generating second-harmonic vortex beams from a fundamental Gaussian wave, which has versatile applications in optical trapping and materials engineering. However, previous studies mainly focused on the normal incidence of the fundamental Gaussian beam, resulting in symmetric emissions of the second-harmonic vortices. Here we present an experimental study on second-harmonic vortex generation in periodic fork-shaped gratings at oblique incidence, in comparison with the case of normal incidence. More quasi-phase-matching resonant wavelengths have been observed at oblique incidence, and the second-harmonic emissions become asymmetric against the incident beam. These results agree well with theoretic explanations. The oblique incidence of the fundamental wave is also used for the generation of second-harmonic Bessel beams with uniform azimuthal intensity distributions. Our study is important for a deeper understanding of nonlinear interactions in a 3D periodic medium. It also paves the way toward achieving high-quality structured beams at new frequencies, which is important for manipulation of the orbital angular momentum of light.

生物组织散射引起的光学像差限制了光学系统的成像性能。本文研究了基于间接波前整形的近红外二区荧光共聚焦成像技术。首先，制备了高效率近红外二区荧光探针，降低该波段生物组织的散射有助于实现高对比度的活体组织成像。其次，研究了基于间接波前测量的自适应光学方法，将间接波前整形技术应用于激光扫描共聚焦显微系统中，以实现对生物组织引起的光学像差的测量与补偿，获得生物组织的高信噪比成像。最后，对基于间接波前整形的近红外二区荧光共聚焦成像系统开展了相关实验。实验结果表明，本系统对空气平板、散射介质和小鼠颅骨等产生的像差具有良好的补偿效果，最终信号强度较初始值分别提升了1.47、1.95和2.85倍，显著提升了最终的成像质量。

In this Letter, we propose and experimentally demonstrate a lens-free wavefront shaping method that utilizes synchronized signal block beam alignment and a genetic algorithm (SSBGA) for a diffuse non-line-of-sight (NLOS) visible light communication (VLC) system. The proposed method effectively controls the position and mobility of visible light beams by partitioning spatial light modulator pixels and manipulating beams to converge at distinct spatial positions, thereby enhancing wavefront shaping efficiency, which achieves a significant 23.9 dB optical power enhancement at +2mm offset, surpassing the lens-based continuous sequence (CS) scheme by 21.7 dB. At +40° angle, the improvement reaches up to 11.8 dB and 16.8 dB compared to the results with and without lens-based CS, respectively. A maximum rate of 5.16 Gbps is successfully achieved using bit-power loading discrete multi-tone (DMT) modulation and the proposed SSBGA in an NLOS VLC system, which outperforms the lens-based CS by 1.07 Gbps and obtains a power saving of 55.6% during the transmission at 4 Gbps. To the best of our knowledge, this is the first time that high-speed communication has been realized in an NLOS VLC system without a lens.

The ever-increasing demand for training and inferring with larger machine-learning models requires more efficient hardware solutions due to limitations such as power dissipation and scalability. Optics is a promising contender for providing lower power computation, since light propagation through a nonabsorbing medium is a lossless operation. However, to carry out useful and efficient computations with light, generating and controlling nonlinearity optically is a necessity that is still elusive. Multimode fibers (MMFs) have been shown that they can provide nonlinear effects with microwatts of average power while maintaining parallelism and low loss. We propose an optical neural network architecture that performs nonlinear optical computation by controlling the propagation of ultrashort pulses in MMF by wavefront shaping. With a surrogate model, optimal sets of parameters are found to program this optical computer for different tasks with minimal utilization of an electronic computer. We show a remarkable decrease of 97% in the number of model parameters, which leads to an overall 99% digital operation reduction compared to an equivalently performing digital neural network. We further demonstrate that a fully optical implementation can also be performed with competitive accuracies.

A fully automatic fail-safe beam shaping system based on a liquid crystal on a silicon spatial light modulator has been implemented in the high-energy kilowatt-average-power nanosecond laser system Bivoj. The shaping system corrects for gain nonuniformity and wavefront aberrations of the front-end of the system. The beam intensity profile and the wavefront at the output of the front-end were successfully improved by shaping. The beam homogeneity defined by the beam quality parameters was improved two to three times. The root-mean-square value of the wavefront was improved more than 10 times. Consequently, the shaped beam from the second preamplifier led to improvement of the beam profile at the output of the first main cryo-amplifier. The shaping system is also capable of creating nonordinary beam shapes, imprinting cross-references into the beam, or masking certain parts of the beam.

Transmission matrix (TM) allows light control through complex media, such as multimode fibers (MMFs), gaining great attention in areas, such as biophotonics, over the past decade. Efforts have been taken to retrieve a complex-valued TM directly from intensity measurements with several representative phase-retrieval algorithms, which still see limitations of slow or suboptimum recovery, especially under noisy environments. Here, we propose a modified nonconvex optimization approach. Through numerical evaluations, it shows that the optimum focusing efficiency is approached with less running time or sampling ratio. The comparative tests under different signal-to-noise levels further indicate its improved robustness. Experimentally, the superior focusing performance of our algorithm is collectively validated by single- and multispot focusing; especially with a sampling ratio of 8, it achieves a 93.6% efficiency of the gold-standard holography method. Based on the recovered TM, image transmission through an MMF is realized with high fidelity. Due to parallel operation and GPU acceleration, our nonconvex approach retrieves a 8685 × 1024 TM (sampling ratio is 8) with 42.3 s on average on a regular computer. The proposed method provides optimum efficiency and fast execution for TM retrieval that avoids the need for an external reference beam, which will facilitate applications of deep-tissue optical imaging, manipulation, and treatment.