Enhancing IoT security with threshold switching-based unified security primitives.

In the era of big data and Internet of Things (IoT), the surge in data volume and widespread interconnectivity of devices make hardware a key node in a vast network ecosystem. Developing state-of-the-art secure IoT hardware is necessary and critical. In this paper, we successfully realized a unified security hardware integrating Physically unclonable function (PUF) and true random number generator (TRNG) functions by constructing a 32× 32 1T1R array based on threshold switching (TS) memristor cells. Through detailed analysis of the chemical properties of FeOfilms and the electrical properties of TS memristors, we verified that the prepared TS devices have good cyclic stability and randomness. Leveraging these characteristics, we implemented a PUF system and effectively deployed a TRNG relying on the CTR_DRBG algorithm. Notably, the generated PUF exhibits a Hamming Weight of 0.508 with the standard deviation of 0.062, along with an Intra-Hamming Distance of 0.00012 with the standard deviation of 0.142 and an Inter-Hamming Distance of 0.496 with the standard deviation of 0.073. Moreover, we conducted a baking test on the samples for 100 h at each temperature point, with intervals of 25 °C within the 25 °C-125 °C range. The resulting bit-error rate (BER) of the generated PUF remains below 1.5%, which clearly attests to the PUF's remarkable uniformity, uniqueness, and robustness. Subsequently, we subjected the generated random numbers to the National Institute of Standards and Technology-900 test. The results revealed that the-values of all test items exceeded 0.01, a strong indication of the outstanding randomness of the generated random numbers. Compared with other similar works, the energy consumption per bit is reduced by more than 30%, and the rate of generating random numbers is increased by more than 20%. The proposed unified security hardware not only demonstrates the potential application of TS memristor in hardware security, but also provides new ideas for solving the integration and energy efficiency problems of security hardware in IoT devices.