N23A-T008   TITLE: Ultra-fast Full-Wave Photonic Simulation and Optimization

OUSD (R&E) CRITICAL TECHNOLOGY AREA(S): Artificial Intelligence (AI)/Machine Learning (ML); General Warfighting Requirements (GWR)

OBJECTIVE: Develop novel ultra-fast simulation technology capable of speeding up full-wave electrodynamic simulation by 1,000 times. This ultra-fast simulation technology will be used to develop long-wave infrared (LWIR) lens that are 100 times thinner and lighter than today�s state-of-the-art LWIR optics and with fully customizable aperture sizes and focal length.

DESCRIPTION: Flat metamaterial optical devices provide a unique opportunity for producing compact and high-performance components for manipulation of light [Ref 1]. Such devices may be constructed through nanofabrication on a planar substrate, providing the possibility of replacing traditional diffractive components that are versatile for various defense and civilian applications [Ref 2]. Flat optics is particularly interesting in the long-wave infrared region because the lack of high-quality imaging optics. Traditional lenses are costly and bulky and possess limited aperture size in the infrared region. Developing integrated imaging optics in LWIR is of great interest especially for tactical surveillance and reconnaissance in this special range [Ref 3].

Metamaterial optical device is a 2D array of dielectric structures that is used to focus transmitted light to a single position directly in front of the device. Typically, these structures are emulated by simulating each dielectric unit cell individually to compute a phase and amplitude transmittance for each cell. While this approach makes for an approximation of the overall device performance, it would be useful to be able to simulate the entire device as a whole to capture the complete physical characteristics of the device structure. However, a simulation of this scale requires several hours or days to perform with a conventional CPU-based finite-difference time-domain (FDTD) method.

To get around the lack of sufficient computational power and to shorten the computation time, the typical compromising approach is to assume a constant phase and amplitude response from each cell [Ref 4], which are used to approximate electromagnetic fields in front of the device. This approximation leads to undesirably less accurate predictions [Ref 5], resulting in sub-optimal device performance. Various techniques, such as the optimization technique in Ref 6, have been attempted to correct for these errors and yet none have been completely successful.

Addressing the very time consuming computation and the accuracy of the modeling and simulation issues simultaneously for this type of complex optics is therefore paramount for designing and fabricating compact optics for LWIR imaging systems with stringent high-performance requirements. This topic seeks to exploit emerging computing hardware to develop ultra-fast computation algorithms to accelerate accurate simulation and uncompromising optimization of flat optics in the LWIR wavelength range. This topic seeks development of FDTD simulation that is at least 1,000 times faster than today�s open source and commercially available FDTD solutions. The developed ultra-high-speed FDTD solution is then utilized to model and simulate metalenses that reduce the computation time�by up to 1,000 times�relative to conventional FDTD methods [Ref 8]. The simulation solution should be able to simulate metalenses of multiple centimeters in optical aperture with optimization of imaging figures of merit such as focusing efficiency and Strehl ratio with respect to the unit cell design parameters. The metalens� focusing efficiency should be higher than 95% over a broad wavelength range from 8�12 microns, and at least 100 times lighter and thinner than traditional state-of-the-art LWIR imaging optics [Ref 8]. Traditional FDTD [Ref 7] would take tens of hours to simulate the abovementioned structure. With the successful development of this algorithm, the same simulation should finish in about one minute.

PHASE I: Determine optimal hardware platform and computational resources required for centimeter-scale simulation of metamaterial lens. Identify both the starting device design based on material and geometry designs from previous experimental works, as well as the figures of merit to optimize and measure experimentally. Demonstrate feasibility and begin development of a working prototype of the software, to include demonstration of a near field calculation of 5 cm x 5 cm flat optics system with a simulation speed improvement of over 1,000 times when compared to CPU-based FDTD. Show consistency between these simulations and experimental demonstrations of the flat metamaterial optics in LWIR. The Phase I effort will include prototype plans to be developed under Phase II.

PHASE II: Develop, demonstrate, and validate the prototype. Incorporate adjoint-based optimization capabilities into the solver. Using adjoint-enabled gradient-based optimization, produce a prototype structure that maximizes the focusing efficiency over the spectral wavelength of interest using about 100 iterations. Demonstrate a significant improvement in device performance that surpasses that of conventional LWIR imaging systems.

PHASE III DUAL USE APPLICATIONS: Transition the technology for Government use. Further develop and refine the design of the simulation software and assist in adapting the simulation software for optimization designs of various photonic devices.

The commercial sectors such as the optics and camera industries can benefit from the reduction of the individual simulation time from several hours to seconds, meaning more parameters can be scanned and at greater resolution, allowing one to simulate structures at a larger scale, enabling full 3D modeling and simulation of an entire device instead of components, and running many jobs simultaneously in the cloud.

REFERENCES:

1.       Yu, N., & Capasso, F. (2014). Flat optics with designer metasurfaces. Nature materials, 13(2), 139-150. https://doi.org/10.1038/nmat3839

2.       Khorasaninejad, M., & Capasso, F. (2017). Metalenses: Versatile multifunctional photonic components. Science, 358(6367). https://doi.org/10.1126/science.aam8100

3.       Zuo, H., Choi, D. -Y., Gai, X., Ma, P., Xu, L., Neshev, D. N., Zhang, B., & Luther-Davies, B. (2017). High-efficiency all-dielectric metalenses for mid-infrared imaging. Advanced Optical Materials, 5(23), 1700585. https://doi.org/10.1102/adom.201800585

4.       Shrestha, S., Overvig, A. C., Lu, M., Stein, A., & Yu, N. (2018). Broadband achromatic dielectric metalenses. Light: Science & Applications, 7(1), 1-11. https://doi.org/10.1038/s41377-018-0078-x

5.       Chung, H., & Miller, O. D. (2020). High-NA achromatic metalenses by inverse design. Optics express, 28(5), 6945-6965. https://doi.org/10.1364/OE.385440

6.       Lin, Z., & Johnson, S. G. (2019). Overlapping domains for topology optimization of large-area metasurfaces. Optics express, 27(22), 32445-32453. https://www.osapublishing.org/DirectPDFAccess/58EECB62-F0FB-4B4F-89A74FA8011960D9_422568/oe-27-22-32445.pdf?da=1&id=422568&seq=0&mobile=no

7.       Warren, C., Giannopoulos, A., Gray, A., Giannakis, I., Patterson, A., Wetter, L., & Hamrah, A. (2019). A CUDA-based GPU engine for gprMax: Open source FDTD electromagnetic simulation software. Computer Physics Communications, 237, 208-218. https://doi.org/10.1016/j.cpc.2018.11.007

8.       Hughes, T. W., Minkov, M., Williamson, I. A., & Fan, S. (2018). Adjoint method and inverse design for nonlinear nanophotonic devices. ACS Photonics, 5(12), 4781-4787. https://pubs.acs.org/doi/pdf/10.1021/acsphotonics.8b01522#

 

KEYWORDS: Full-wave photonic simulation; Design optimization; Computational speed improvement; Metalens; Metamaterials; Lens

---