Jump to main content
Jump to site search

Nanoscale Diffusive Memristor Crossbars as Physical Unclonable Functions


Physical unclonable functions have emerged as promising hardware security primitives for device authentication and key generation in the era of Internet of Things. Here we report novel physical unclonable functions built upon crossbars of nanoscale diffusive memristors that translate the stochastic distribution of Ag clusters in a SiO2 matrix into a random binary bit map that serves as a device fingerprint. The random dispersion of Ag leads to an uneven number of clusters at each cross-point, which in turn results in stochastic switching probability of the Ag:SiO2 diffusive memristors in an array. The randomness of the dispersion is a barrier to fingerprint cloning, and the unique fingerprints of each device are persistent after fabrication. With optimized fabrication procedure, we have maximized the randomness and achieved an inter-class Hamming distance of 50.68 %. We also discovered that bits are not flipping after over 104 s at 400 K, suggesting an intra-class Hamming distance of 0 % and hence superior reliability of our physical unclonable functions. In addition, our diffusive memristor-based physical unclonable functions are easy to fabricate, do not require complicated post processing for digitization, and thus provide new opportunities in hardware security applications.

Back to tab navigation

Publication details

The article was received on 03 Sep 2017, accepted on 10 Jan 2018 and first published on 11 Jan 2018

Article type: Communication
DOI: 10.1039/C7NR06561B
Citation: Nanoscale, 2018, Accepted Manuscript
  •   Request permissions

    Nanoscale Diffusive Memristor Crossbars as Physical Unclonable Functions

    R. Zhang, H. Jiang, Z. Wang, P. Lin, Y. Zhuo, D. Holcomb, D. Zhang, J. J. Yang and Q. Xia, Nanoscale, 2018, Accepted Manuscript , DOI: 10.1039/C7NR06561B

Search articles by author