CMOS Optical PUFs Using Noise-Immune Process-Sensitive Photonic Crystals Incorporating Passive Variations for Robustness

Xuyang Lu, Lingyu Hong, Kaushik Sengupta

Research output: Contribution to journalArticlepeer-review

6 Scopus citations

Abstract

This paper introduces a new design methodology for incorporating process-sensitive optical nanostructures in standard CMOS processes to create robust optical physically unclonable functions (PUFs) realized through an electrical-photonic co-design approach. The passive lithographic variations of lower level metal interconnects are exploited to realize resonant photonic crystals on an array of photodetectors to include variations that are robust to noise processes. The chip is realized in a standard 65-nm CMOS process with no additional post-processing. The addition of the structures increases the coefficient of variation by a factor of 3.5 × compared to only active device variations. This creates extremely robust PUF responses with a native inter-chip Hamming distance (HD) of 49.81% and intra-chip HD of 0.251% with an inter-HD/ intra-HD ratio of 198 × illustrating the reliability of the design. The native intra-HD can be reduced to 0.06% with 17 mV of thresholding with only 4% of the total combinations discarded. To the best our knowledge, this is also the first demonstration of photonic crystals and an optical PUF in CMOS.

Original languageEnglish (US)
Article number8430651
Pages (from-to)2709-2721
Number of pages13
JournalIEEE Journal of Solid-State Circuits
Volume53
Issue number9
DOIs
StatePublished - Sep 2018

All Science Journal Classification (ASJC) codes

  • Electrical and Electronic Engineering

Keywords

  • CMOS
  • CMOS imager
  • Chip identification
  • PUFs
  • optical physically unclonable functions (PUFs)
  • photodetectors
  • photonic crystals
  • process variations

Fingerprint Dive into the research topics of 'CMOS Optical PUFs Using Noise-Immune Process-Sensitive Photonic Crystals Incorporating Passive Variations for Robustness'. Together they form a unique fingerprint.

Cite this