We discuss nanoscale 3D-electronic-photonic-integrated-circuits (3D EPICs) and their applications in future intelligent imaging systems. The talk will be in two parts emphasizing heterogeneous integration of imaging and AI computing capabilities in microsystems. First, we will discuss a new generation of compressive hyperspectral Imaging systems with 3D EPICs exploiting interferometric imaging and compressive sensing techniques to achieve high-resolution hyperspectral imaging at ~100x smaller size, weight, and power compared to the current state of the art. Recent experimental demonstrations achieved successful image reconstructions from interferometric imaging using photonic integrated circuits. Second, we will discuss AI computing systems consisting of neuromorphic computing and von Neumann computing systems. Neuromorphic computing systems include attojoule nanoscale optoelectronic neurons in 3D photonic neural networks to pursue brain-derived computing with self-learning capabilities, while von Neumann computing systems consist of 3D EPIC fine-grain accelerators and memory units interconnected by all-to-all photonic networks.

Advancements in 3D heterogeneously integrated (3DHI) microsystems has the potential to radically change the compute and sensor system design. The dense and intimate integration of multiple chips can dramatically increase the interconnect bandwidth and increase the functionality of sensors and processor chips. The inclusion of integrated photonics in these architectures enables many of these benefits. This talk will share recent DARPA program investments in enabling these architectures. It will also discuss new challenges and opportunities these architectures present for heterogenous integration, photonic integrated circuit and microfabrication manufacturing technology.

HgCdTe-based FPAs that can be used in high neutron radiation environments were designed and fabricated by EPIR, and tests using Fermi National Accelerator Laboratory’s neutron beam confirmed that these FPAs can maintain imaging functionality while exposed to fluxes up to low-1E13 neutron per squared centimeter accumulated neutron exposure. Monte Carlo N-Particle (MCNP) simulations were used to find that the energy deposited into the HgCdTe FPA can come from not only directly impinging neutrons but also scattered neutrons and subsequently generated protons, electrons and photons, confirming that our neutron-hardened designs are also hardened against other high energy particles. To mitigate radiation damage, we redesigned the optical system of the camera using modeling and simulation by utilizing MCNP code during our camera design. By properly choosing mirror substrate material and coating as well as the corresponding optical system and the camera design, we can filter out harmful radiation flux while still collecting the MWIR signal with high efficiency, thereby significantly reducing camera and image system performance degeneration under high-energy high-flux neutron beams.

The past few years have seen integrated photonic devices beginning to outperform their bulk optics counterparts. Such advances, combined with other advantages such as superior environmental stability and system scalability offered by the solid-state nature of photonic integrated circuits, offer the potential for optical microsystems that not only miniaturize optical technologies, but further achieve functionality that is impossible or impractical at the macroscopic scale. This talk will present case studies of such developments from current DARPA programs, describe opportunities to achieve capability advantages in future precision sensing research program areas, and explore the technical challenges inherent in addressing these opportunities.

High-performance systems are increasingly bottlenecked by the energy and communications costs of interconnecting numerous compute and memory resources. Integrated silicon photonics offer the opportunity of embedding optical connectivity that directly delivers high off-chip communication bandwidth densities with low power consumption. Our recent work has shown how integrated silicon photonics with comb-driven dense wavelength-division multiplexing can scale to realize Pb/s chip escape bandwidths with sub-picojoule/bit energy consumption. This talk will discuss the integrated photonic link as well as the multi-chip packaging implemented to realize energy efficient high bandwidth density chip IO.

Reconfigurable optical metasurfaces are rapidly emerging as a major frontier in photonics research, development, and applications. They promise compact, lightweight, and energy-efficient reconfigurable optical systems with unprecedented performance and functions that can be dynamically defined on-demand. Space applications represent an emerging area in which these characteristics are highly prized. The ability to dynamically tune optical functions through selective modulation of electromagnetic waves is crucial to the advancement of a variety of sensing applications, from imaging spectrometers to light detection and ranging (LiDAR). This presentation introduces a reconfigurable metasurface optic project led by a research team at NASA Langley Research Center (NASA LaRC) since 2019. It covers advances in phase change material-based reconfigurable optics, performance data on reliability enhancement of photonic devices, image sensing system architectures, and mission concepts enabled through these advances.

Reconfigurable or tunable metasurfaces are of great interest for adaptive multi-modal sensing and imaging. Phase change materials (PCMs) have been recently explored for reconfigurable optical devices such as spatial light modulators and metasurfaces. In this talk, we will present large-scale reconfigurable infrared metasurfaces consisting of greater than 30 million pixelated optical antennas integrated with PCM mesa structures over a 1cm x 1cm area.

Optical metasurfaces are artificially structured surfaces that allow abrupt modification of the properties of light at sub-wavelength spatial scales. Due to the intimate control over light fields that they enable and also because of their ultra-thin character, they are finding increasing applications to simplify and miniaturize optical systems. In this regard, their use in imaging, sensing and metrology systems is among the most prominent ones. In this talk, we will present some of our recent results in the use of optical metasurfaces to achieve ultra-wide-angle, multispectral imaging and 3D metrology, and also share our recent development on their use for hyperspectral imaging in the MWIR for space applications. In all these topics, we will present the fundamental physical concepts behind the unique properties of these meta-optical devices, show pathways for their mass manufacturing and share thoughts on possible future avenues for this emerging technology.

This talk is focused on new material and device platforms for realization of reconfigurable metaphotonic devices using materials with a large variation of their optical properties through phase transformation. Volatile and nonvolatile phase-change materials as well as polymer-based materials will be covered. Design, fabrication, and application of this platform for state-of-the-art devices and systems will also be discussed.

Nanophotonic devices have gained attention as promising solutions for all-optical image processing. These filters are of subwavelength size and carry potential to address limitations of current image processing technology including processing speed, energy requirements as well as size. We present results demonstrating the use of thin-film absorbers and metasurfaces to real-time detection of edges in images and the visualisation of phase objects including human cancer cells. We discuss the extension of these approaches to implement tuneable devices using phase-change materials and graphene. These approaches have potential for integration into ultra-compact mobile medical diagnostic tools, as well as remote sensing systems.

Machine-learning algorithms are powerful tools in developing reliable models to relate the design space of a nanophotonic structure to its response space. They can be used not only to simplify the inverse design problem but also to provide valuable insight about the physics of light-matter interaction. This talk will provide a new approach through combining manifold-learning algorithms for reducing the dimensionality of the problem with metric-learning techniques for more insightful mapping of the input-output relation to the dimensionality-reduced (or the latent) space. In addition to covering the fundamental properties of the presented algorithms, their applications to both the inverse design and the knowledge discovery in state-of-the-art metaphotonic structures will be discussed.

The DARPA Optomechanical Thermal Imaging (OpTIm) program seeks to develop a novel modality of low-SWaP quantum-limited infrared detection. Building upon proof-of-concept studies and detailed noise analyses, the OpTIm device concept amalgamates high performance optomechanical sensors, quantum-limited all-optical readout techniques, and spectrally resolved metamaterial-based IR absorbers to realize several order-of-magnitude improvements in both IR sensitivity and detection speed. I will give a brief overview of the OpTIm program concept, the over-arching goals of this program, and a summary of innovations that have resulted from the program thus far. I will also describe the numerous technological applications and opportunities that are engendered by the novel capabilities of OpTIm imaging systems.

In this talk we review fundamental concepts in chip-scale cavity optomechanics, laser cooling and driven oscillator efforts, in the silicon platform compatible with microelectronics. We describe planar-integrated oscillators of large optomechanical transduction, through a photonic crystal slot cavity for deeply-sub-wavelength [≈ 0.1(λ/n)3] electromagnetic localization. Driven with below-mW powers, the optomechanical sensing pixel has dispersive couplings g*/2π up to 783 kHz, large temperature coefficients with 0.44% frequency shift per Kelvin, and distinguishable frequency shifts (ΔΩ_m) for the mechanical oscillator under infrared-radiation at room temperature. We examine thermodynamical and fundamental bounds on the laser-driven oscillator and RF-readout, integrating into higher technology readiness level prototypes.

We describe a new type of high-performance IR sensors based on ultrathin lithium niobate (LN) resonant mechanical resonators. Optimal IR absorption will be achieved by nanometer-thin metal films that are impedance matched to free space; narrowband spectral response will be achieved by metamaterials patterning. LN provides many attractive features as compared to existing IR sensor materials, including strong piezoelectric coupling, large temperature coefficients, and favorable thermal conductance and capacitance. Additionally, recent advances in LN-on-insulator (LNOI) technology now permit exceptional film thickness and stress control, making it feasible to create ultrathin phononic structures that rival SiN-based IR sensors – but promise greatly enhanced sensitivity and temporal resolution.