Exploiting Device-Specific Noise Signatures as Hardware Authentication Fingerprints on Near-Term Quantum Processors

Physically Unclonable Functions (PUFs) are a cornerstone security primitive in classical hardware
authentication, exploiting manufacturing-induced physical variations to generate device-unique,
unclonable identifiers. In this work, we propose and experimentally evaluate the first Quantum PUF (QPUF) protocol implemented on a multi-device quantum simulation framework, using PKTRON v3.7.3
with four distinct virtual quantum processors: PK Falcon 27Q, PK Eagle 65Q, PK IonTrap 16Q, and PK
NoisyLab 8Q. The Q-PUF protocol operates by executing a library of eight standardized challenge
circuits on each device and treating the resulting noisy probability distributions as device fingerprints.
We conduct seven systematic experimental phases covering enrollment, inter-device distinguishability,
intra-device stability, authentication, impersonation attack resistance, noise deviation analysis, and perchallenge discriminability. Our results reveal a fundamental tension at the core of Q-PUF design: while
all four devices exhibit high intra-device stability (mean cosine similarity > 0.997 across all devices), the
inter-device cosine similarity is comparably high (> 0.999), rendering cosine-similarity-based
authentication insufficient at a fixed threshold of 0.90. We identify this as the Q-PUF Distinguishability
Problem — the noise profiles of near-term virtual devices are sufficiently similar that standard similarity
metrics conflate intra-device reproducibility with inter-device proximity. We analyze the per-challenge
discriminability of eight circuit archetypes, finding that rotation-based circuits (C3_RotLadder) achieve
the highest discriminability score (0.3515), while deterministic circuits (C4_AltX) contribute zero
discriminability. These findings define concrete design requirements for viable Q-PUF systems and
motivate adaptive authentication thresholds, higher-dimensional fingerprint spaces, and physically
diverse hardware platforms as necessary conditions for practical Q-PUF deployment.