ECE’s Guangyu Xu and Team Report on Pioneering Reconfigurable Silicon Device for Cybersecurity

Content

The ubiquitous presence of interconnected devices in the era of Internet-of-Things (IoT) highlights the pressing need to prevent unauthorized access of sensitive data and computing algorithms. Now, to address this issue, Guangyu Xu – an associate professor in the UMass Amherst Electrical and Computer Engineering (ECE) Department and an adjunct of the Biomedical Engineering Department – and his collaborators have published a trailblazing paper in IEEE Electron Device Letters reporting their research on a promising new cryptographic device. IEEE Electron Device Letters is a prestigious journal published by the Institute of Electrical and Electronics Engineers.

Their paper describes cutting-edge research on a dual-gate silicon photodiode as a new type of optoelectronic physical unclonable function (PUF), which offers its users high reconfigurability, high resistance to adversarial attack, all-silicon construction, and low overhead.

Besides Xu, the other researchers and authors of the paper are ECE Ph.D. students Matthew Donnelly, Wen Liang, and Meiyue Zhang and ECE Distinguished Professor Lixin Gao. Their collaborative research is supported in part by awards from the National Science Foundation.

According to Xu, the research reported in this paper represents a new direction in his laboratory, which in the past has largely focused on integrated optoelectronic arrays for high-precision cell interfacing and biosensing. This work marked the beginning of his new endeavor in the area of hardware security, leveraging his cohesive collaboration with Gao, a renowned authority on privacy-preserving IoT networks.

Leveraging its high-dimensional photo response obtained by gate sweeping, this new PUF is noted to generate security labels that are uniform, unique, reliable, and resistant to machine-learning-based malign attacks. Furthermore, its high reconfigurability leads to high-capacity keys that can effectively mask user images from attackers, well-poised to boost cybersecurity in resource-limited edge devices.

Xu and his team believe that the idea of repurposing photodiodes towards security hardware is offering a CMOS-compatible, attack-resistant PUF technology, whose large number of reconfigurable states (45) benefits privacy-preserving IoT applications such as anti-counterfeiting image classification.

The research team concludes that “Overall, our high-performing, all-silicon, PUF technology holds promise as a solution for low-overhead IoT edge cryptosystems.” 

Xu, who is a Senior Member of the IEEE, is a well-known authority on research that involves neuro-engineering, precision medicine, nanotechnology, bioinformatics, machine vision, and more. Gao, a Fellow of the IEEE and the Association for Computing Machinery, is a widely recognized expert on Internet routing, cloud computing, distributed frameworks for machine learning, and privacy-preserving data analytics. (March 2026)