High-speed spatial modulation of light is the key technology in various applications, such as optical communications, imaging through scattering media, video projection, pulse shaping, and beam steering, in which spatial light modulators (SLMs) are the underpinning devices. Conventional SLMs, such as liquid crystal (LC), digital micromirror device (DMD), and micro-electro-mechanical system (MEMS) ones, operate at a typical speed on the order of several kilohertz as limited by the slow response of the pixels. Achieving high-speed spatial modulation is still challenging and highly desired. Here, we demonstrate a one-dimensional (1D) high-speed programmable spatial light modulator based on the electro-optic effect in lithium niobate thin film, which achieves a low driving voltage of 10 V and an overall high-speed modulation speed of 5 MHz. Furthermore, we transfer an image by using parallel data transmission based on the proposed lithium niobate SLM as a proof-of-principle demonstration. Our device exhibits improved performance over traditional SLMs and opens new avenues for future high-speed and real-time applications, such as light detection and ranging (LiDAR), pulse shaping, and beam steering.

Pulse shaping plays a key role in various applications of ultrafast lasers, such as optical communications, laser micromachining, microscopy, and quantum coherent control. Conventional pulse shaping devices based on liquid crystal spatial light modulators (LCSLMs) or digital micromirror devices (DMDs) only have the shaping speed of several hertz to kilohertz, which is not suitable for applications requiring a high-speed response. Here, we demonstrate a high-speed programmable lithium niobate spatial light modulator (LNSLM) with 128 individual modulation channels and a modulation speed that can reach 1 MHz. Then we establish a high-speed LNSLM-based Fourier-transform (FT) pulse shaper to realize high-speed pulse shaping, and the update rate can reach 350 kHz, only limited by the electric circuit. The proposed high-speed pulse shaper scheme opens new avenues for future applications of ultrafast science, such as microscopic imaging, interaction between light and matter, and spectroscopy.

No abstract

The wavemeter is an important instrument for spectrum analysis, widely used in spectral calibration, remote sensing, atomic physics, and high-precision metrology. However, near-infrared (NIR) wavemeters require infrared-sensitive detectors that are expensive and less sensitive compared to silicon-based visible light detectors. To circumvent these limitations, we propose an NIR speckle wavemeter based on nonlinear frequency conversion. We combine a scattering medium and the deep learning technique to invert the nonlinear mapping of the NIR wavelength and speckles in the visible wave band. With the outstanding performance of deep learning, a high-precision wavelength resolution of 1 pm is achievable in our experiment. We further demonstrate the robustness of our system and show that the recognition of power parameters and multi-spectral lines is also feasible. The proposed method offers a convenient and flexible way to measure NIR light, and it offers the possibility of cost reduction in miniaturized wavemeter systems.