Fig. 1: Spectral PUFs for authentication of goods and verification of individuals: a schematic overview of the work.

First, the laser-induced fabrication of Au/Si structures yields submicroparticles with distinct intricate internal configurations. The random nanoscale structural variations of these particles greatly contribute to their multi-resonant nature, producing bright, broadband PL spectra with unique characteristics. Leveraging this spectral randomness, we propose two anti-counterfeiting label models for quick and reliable authentication. The first model utilizes PL mapping and spectrum decomposition, and the second one employs direct PL imaging of printed pixel patterns under unfocused laser beam excitation. Both approaches use weak PUFs, where each label’s signature serves as a key stored in a static database (DB). Also, we note that the time required to measure all possible PL spectra from one laser-printed microstructure sufficiently surpasses its fabrication time, enabling the realization of a strong PUF. Building upon this concept, we introduce a one-time password verification token. Our protocol utilizes individual PL spectra, measured from randomly selected points within the printed array of microstructures, as unique access keys. To ensure high-level security, each key is valid for a single verification attempt. A high-accuracy machine learning model facilitates robust comparison of PL spectra. For a deeper understanding of the concepts illustrated in this figure, please refer to the main text of the manuscript.