Solid-state beam steering is the key to realize miniature, mass-producible LIDAR (Light Detection And Ranging) and freespace communication systems without using any moving parts. The huge power consumption required in solid-state beam steering, however, prevents this technology from further scaling. Here we show two different approaches to enable lowpower solid-state beam steering. In the first approach, we use spatial-mode multiplexing to reduce the power consumption of the phase shifters in a large-scale optical phased array. We show an improvement of phase shifter power consumption by nearly 9 times, without sacrificing optical bandwidth or operation speed. Using this approach, we demonstrate 2D beam steering with a silicon photonic phased array containing 512 actively controlled elements. This phased array consumes only 1.9 W of power while steering over a 70° × 6° field of view. This power consumption is at least an order of magnitude lower compared to other demonstrated large-scale active phased arrays. In the second approach, we achieve 2D beam steering with a switchable emitter array and a metalens that collimates the emitted light. The power consumption of this approach scales logarithmically with the number of emitters and therefore favors large-scale systems. This approach allows straightforward feedback control and better robustness to environmental temperature change. Our approaches demonstrate a path forward to build truly scalable beam steering devices.
High spatiotemporal resolution deep-brain optical excitation for optogenetics would enable activation of specific neural populations and in-depth study of neural circuits. Conventionally, a single fiber is used to flood light into a large area of the brain with limited resolution. The scalability of silicon photonics could enable neural excitation over large areas with single-cell resolution similar to electrical probes. However, active control of these optical circuits has yet to be demonstrated for optogenetics.
Here we demonstrate the first active integrated optical switch for neural excitation at 473 nm, enabling control of multiple beams for deep-brain neural stimulation. Using a silicon nitride waveguide platform, we develop a cascaded Mach-Zehnder interferometer (MZI) network located outside the brain to direct light to 8 different grating emitters located at the tip of the neural probe. We use integrated platinum microheaters to induce a local thermo-optic phase shift in the MZI to control the switch output. We measure an ON/OFF extinction ratio of >8dB for a single switch and a switching speed of 20 microseconds. We characterize the optical output of the switch by imaging its excitation of fluorescent dye.
Finally, we demonstrate in vivo single-neuron optical activation from different grating emitters using a fully packaged device inserted into a mouse brain. Directly activated neurons showed robust spike firing activities with low first-spike latency and small jitter. Active switching on a nanophotonic platform is necessary for eventually controlling highly-multiplexed reconfigurable optical circuits, enabling high-resolution optical stimulation in deep-brain regions.
GRIN (Graded index) lens have revolutionized micro endoscopy enabling deep tissue imaging with high resolution. The challenges of traditional GRIN lenses are their large size (when compared with the field of view) and their limited resolution. This is because of the relatively weak NA in standard graded index lenses.
Here we introduce a novel micro-needle platform for endoscopy with much higher resolution than traditional GRIN lenses and a FOV that corresponds to the whole cross section of the needle. The platform is based on polymeric (SU-8) waveguide integrated with a microlens micro fabricated on a silicon substrate using a unique molding process. Due to the high index of refraction of the material the NA of the needle is much higher than traditional GRIN lenses. We tested the probe in a fluorescent dye solution (19.6 µM Alexa Flour 647 solution) and measured a numerical aperture of 0.25, focal length of about 175 µm and minimal spot size of about 1.6 µm. We show that the platform can image a sample with the field of view corresponding to the cross sectional area of the waveguide (80x100 µm2). The waveguide size can in principle be modified to vary size of the imaging field of view. This demonstration, combined with our previous work demonstrating our ability to implant the high NA needle in a live animal, shows that the proposed system can be used for deep tissue imaging with very high resolution and high field of view.