Physical unclonable functions based on chaotic vertical-cavity surface-emitting lasers for dynamic authentication

This ZIP archive contains the source data used to generate supplementary figures and analyses in the manuscript. A description of each file is provided below: Fig1d_Opt_Spectrum_comparison.xlsx Optical spectra comparison under different operating biases. Fig1efg_Performance_comparison.xlsx Performance metrics (e.g., classification accuracy, robustness) under various model and configuration settings. Fig2b_Histogram.xlsx Histogram plots of key distribution metrics derived from chaotic outputs. Fig2cdfij_All_correlation.xlsx Correlation matrices across multiple devices and conditions to assess independence and variability. Fig2e_Device-to-device_FFT-spectrum_comparison.xlsx FFT-based spectral comparisons between different devices under the same bias. Fig3bc_Algorithm.xlsx Detailed steps and intermediate results of the AI-based authentication algorithm. Fig4bcdf_Generative_framework.xlsx Evaluation results of the generative model (e.g., VAE) used in spoofing and reconstruction analysis. Supplementary_Attack.xlsx Data from attack simulations including model predictions, confidence levels, and spoofing success rates. Supplementary_CRP_Stability&separability.xlsx Key stability and separability analysis across repeated CRPs and challenge configurations. Supplementary_FFT-variations_different_env.xlsx FFT spectra under environmental variations (e.g., temperature drift, power cycling). Supplementary_Hmin_LE.csv Minimum entropy and Lyapunov exponent estimates for different devices and bias conditions. Supplementary_inter-&intra-key_HD.csv Inter- and intra-device Hamming distance (HD) distributions of generated keys across multiple conditions.