Thermal Stress Effects on Physical Entropy in Quantum and True Random Number Generators

This work examines how thermal stress (0 ºC - 60 ºC) affects the physical entropy of commercial Quantum (QRNG) and True (TRNG) Random Number Generators. Using the NIST SP 800-90B test suite, we show that elevated and extreme temperatures induce hardware-level structural asymmetries. With the macroscopic chi-square test, we find that TRNG deteriorate approximately linearly with temperature, while QRNG displays more complex but effectively temperature-invariant behavior due to active hardware compensation. We further show that fixed-rate cryptographic extractors can mask these physical degradations, silently weakening information-theoretic security. These findings highlight the need for adaptive and environment-sensitive randomness extraction to ensure robust cryptographic security under realistic operating conditions.