In the field of engineering and materials science, the identification of concealed structures or imperfections within materials is of paramount importance.
Illustration of a single-pixel diffractive terahertz sensor, which can detect hidden defects within 3D samples using a snapshot illumination, along with the photo of its 3D-printed prototype. Image Credit: Ozcan Lab @ University of California Los Angeles
Conventional terahertz imaging systems leverage the unique property of terahertz waves to penetrate materials that are visually opaque.
These systems have been developed to uncover the internal structures of various materials, offering significant advantages in a wide range of applications, including industrial quality control, security screening, biomedicine, and defense.
However, most existing terahertz imaging setups suffer from limited throughput, bulkiness, and the need for raster scanning to capture images of concealed features.
To revolutionize this approach, researchers from the UCLA Samueli School of Engineering and the California NanoSystems Institute have devised an innovative terahertz sensor.
This sensor can swiftly identify hidden defects or objects within a target sample volume using a single-pixel spectroscopic terahertz detector.
Instead of the traditional method involving point-by-point scanning and digital image formation, this sensor examines the entire volume of the test sample illuminated with terahertz radiation in a single snapshot without the need to generate or digitally process an image of the sample.
Led by Dr. Aydogan Ozcan, Chancellor's Professor of Electrical & Computer Engineering, and Dr. Mona Jarrahi, the Northrop Grumman Endowed Chair at UCLA, this sensor functions as an all-optical processor, adept at detecting and categorizing unexpected wave sources resulting from diffraction through concealed defects.
The study has been published in the journal Nature Communications.
It is a shift in how we view and harness terahertz imaging and sensing as we move away from traditional methods toward more efficient, AI-driven, all-optical sensing systems.
Dr. Aydogan Ozcan, Associate Director, California NanoSystems Institute, UCLA Engineering Institute for Technology Advancement
The novel sensor is constructed with a sequence of diffractive layers, which are automatically fine-tuned through the application of deep learning algorithms.
Once this optimization process is complete, the layers are translated into a tangible prototype using additive manufacturing techniques like 3D printing. Consequently, the system can engage in all-optical processing, eliminating the cumbersome requirements of raster scanning or digital image acquisition and processing.
It is like the sensor has its own built-in intelligence. Our design comprises several diffractive layers that modify the input terahertz spectrum depending on the presence or absence of hidden structures or defects within materials under test. Think of it as giving our sensor the capability to 'sense and respond' based on what it 'sees' at the speed of light.
Dr. Aydogan Ozcan, Associate Director, California NanoSystems Institute, UCLA Engineering Institute for Technology Advancement
Ozcan made this observation while drawing comparisons with their previous AI-designed optical neural networks. To showcase their innovative idea, the UCLA team manufactured a diffractive terahertz sensor through 3D printing and effectively identified concealed flaws in silicon specimens.
These specimens were comprised of layered wafers, with one layer containing defects and the other masking them. The intelligent system precisely exposed the existence of undisclosed defects of diverse shapes and positions.
The team is confident that their diffractive defect sensor framework can extend its applicability to various wavelengths, including infrared and X-rays. This adaptability opens the door to a wide range of potential applications, spanning manufacturing quality control, security screening, and even cultural heritage preservation.
The straightforward nature, substantial throughput, and cost-effectiveness of this non-imaging approach hold the potential to bring about significant advancements in scenarios where speed, efficiency, and precision are of utmost importance.
Journal Reference
Li, J., et al. (2023) Rapid sensing of hidden objects and defects using a single-pixel diffractive terahertz sensor. Nature Communications. doi.org/10.1038/s41467-023-42554-2.
Source: https://www.ita.ucla.edu/