OUSD (R&E) MODERNIZATION PRIORITY: Microelectronics, AI/ML
TECHNOLOGY AREA(S): Information Systems, Modeling and Simulation Technology
OBJECTIVE: Develop advanced THz-FPA that offer large pixel count, high dynamic range, and high speed over a broad THz frequency range.
DESCRIPTION: Electromagnetic waves in the THz spectral band (roughly covering the 0.1 – 3 THz frequency range) offer unique properties for chemical identification, nondestructive imaging, and remote sensing. However, existing THz devices have not yet provided all the functionalities required to fulfill many of these applications. Although complementary metal–oxide–semiconductor (CMOS) technologies have been offering robust solutions below 1 THz, the high-frequency portion of the THz band still lacks mature devices. For example, most of the THz imaging and spectroscopy systems use single-pixel detectors, which results in a severe tradeoff between the measurement time and field of view. To address this problem, a large pixel count, high dynamic range, high speed, and broadband THz-FPA needs to be developed. The proposed THz-FPA can operate either as a frequency-tunable continuous-wave detector or a broadband-pulsed detector. It should be able to operate over a 1 – 3 THz frequency range while offering more than 30 decibel (dB) dynamic range per pixel. It should have more than 1,000 pixels and a frame rate of at least 1 hertz (Hz). Some anticipated features include developing THz-FPAs by exploring three-dimensional microstructures, smart readout integrated circuits, and processors that incorporate neuromorphic computing and ML to increase the data collection efficiency.
Direct Phase 2 Proposals—that is, proposals that skip Phase I—are being accepted under this topic. Such proposals should describe existing Thz technologies and their challenges, contrast with the proposed effort, and build a prototype with a long dynamic range. The effort should clearly justify the rational for a direct Phase II proposal and identify clear milestones. A direct Phase II proposal must include strong evidence of a verified standard FPA with comparable frequency range.
PHASE I: Demonstrate a proof-of-concept THz-FPA with at least 16 pixels. Show that each pixel of the THz-FPA meets the dynamic range and bandwidth requirements. Introduce a data readout method that can maintain the large dynamic range and broad bandwidth requirements for more than 1,000 pixels and a frame rate of at least 1 Hz. Develop a Phase II plan that includes technology integration, test, and validation with representative structures.
PHASE II: Realize the THz-FPA consisting of at least 1,000 pixels integrated with the read-out circuits. Demonstrate the functionality of the final prototype to take THz images with more than a 30dB dynamic range over a 1 – 3 THz bandwidth in less than 1 second. The prototype system will vary based on the proposed approach, but it may include hardware and software. Develop a technology transition plan and business case assessment.
PHASE III DUAL USE APPLICATIONS: Broadband THz imaging FPA enable sensors for detailed feature and frequency spectrum capture that support several DoD missions, including battlespace target assessment, remote sensing, surveillance in low-visibility conditions, nondestructive material quality control; law enforcement missions to detect illicit drugs and narcotics; and regulatory agencies for detecting toxins in drug, food, and agricultural products.
- C. Posch, T. Serrano-Gotarredona, B. Linares-Barranco, T. Delbruck. Retinomorphic event-based vision sensors: Bioinspired cameras with spiking output. Proceedings of the IEEE 102, 10, 1470–1484, Oct. 2014, doi:10.1109/JPROC.2014.2346153.
- K. Kito, T. Matsui, S. Kidera. Depth-adaptive object identification using terahertz time domain spectroscopic data. IEEE Transactions on Terahertz Science and Technology 11.5 (2021): 598–604.
- B. Limbacher, S. Schoenhuber, M. Wenclawiak, M.A. Kainz, A.M. Andrews, G. Strasser, J. Darmo, K. Unterrainer. Terahertz optical machine learning for object recognition. APL Photonics 5, 126103 (2020). https://doi.org/10.1063/5.0029310.
KEYWORDS: Terahertz focal plane, imaging, optical machine learning, CMOS