Search

Organic electro-optic (OEO) materials have demonstrated promising performance in developing electro-optic phase shifters (EOPS) and modulators compared to their inorganic counterparts. Integration with other devices in a silicon photonic (SiP) process, simple nanofabrication, and temperature/aging robustness remain to be developed for this class of hybrid material platforms. In particular, electro-optic (EO) polymers need an electro-thermal poling method, which has limited their potential and utilization in large-scale SiP. Devices made of paraelectric nematic liquid crystals (PN-LC), another primary type of OEO material, feature a very efficient but slow phase shift mechanism. We present a general-purpose EOPS that applies to various modulator embodiments to address these concerns. Based on that, we report a GHz-fast phase shift in a newly discovered family of OEO, namely ferroelectric nematic liquid crystals (FN-LC), which finally enables liquid crystals to have significant second-order nonlinear optical coefficients and associated Pockels effect. The new material avoids poling issues associated with EO polymers and can pave the way for hybrid silicon-OEO systems with CMOS-foundry compatibility. Furthermore, we propose a finger-loaded, non-slotted waveguide that enhances light-matter interaction, allowing us to achieve DC and AC modulation efficiencies of $\approx$ 0.25 V.mm and 25.7 V.mm, respectively, an on-chip insertion loss of $\approx$ 4 dB, and an EO bandwidth of f$_{-6dB}$ >4.18 GHz. The remaining figures of merit for our poling-free EOPS are equivalent to EO polymer-enabled devices with fewer manufacturing difficulties. We demonstrate an electrically and photonically packaged chip that contains >100 silicon-FN-LC modulators to evaluate the large-scale integration of our poling-free phase shifters and modulators., Comment: 21 pages, 5 figures

Photonics offers a transformative approach to artificial intelligence (AI) and neuromorphic computing by providing low latency, high bandwidth, and energy-efficient computations. Here, we introduce a photonic tensor core processor enabled by time-multiplexed inputs and charge-integrated outputs. This fully integrated processor, comprising only two thin-film lithium niobate (TFLN) modulators, a III-V laser, and a charge-integration photoreceiver, can implement an entire layer of a neural network. It can execute 65 billion operations per second (GOPS) per neuron, including simultaneous weight updates-a hitherto unachieved speed. Our processor stands out from conventional photonic processors, which have static weights set during training, as it supports fast "hardware-in-the-loop" training, and can dynamically adjust the inputs (fan-in) and outputs (fan-out) within a layer, thereby enhancing its versatility. Our processor can perform large-scale dot-product operations with vector dimensions up to 131,072. Furthermore, it successfully classifies (supervised learning) and clusters (unsupervised learning) 112*112-pixel images after "hardware-in-the-loop" training. To handle "hardware-in-the-loop" training for clustering AI tasks, we provide a solution for multiplications involving two negative numbers based on our processor., Comment: 19 pages, 6 figures