Loading...
Please wait, while we are loading the content...
Similar Documents
Schedule of Events 8 : 00 AM-8 : 30 AM Breakfast and Registration 8 : 30 AM-8 : 45 AM Introduction
| Content Provider | Semantic Scholar |
|---|---|
| Author | Nasr, Joseph Elahi, Asim Zhao, Hongfan Anil, Vivek |
| Copyright Year | 2019 |
| Abstract | Recent years have witnessed a rapid growth in graphene based electronic devices and sensors for applications in high resolution displays, wearable electronics, biological and chemical sensors, etc. These high performance, low power graphene devices are critical for various emerging technologies such as the Internet of Things (IoT). However, such technologies also thrive on interconnectivity of electronic components for sharing digital information making them vulnerable to cyber-attacks, tampering, hacking, and other security threats. As such, in-build and on-chip hardware security can greatly benefit the development of graphene based electronics. Here we show that the disorders in carrier transport in graphene field effect transistors (GFETs), originating from mobility fluctuation and random distribution of grain boundaries, defects and impurities can be used to generate a physically unclonable functions (PUFs) for cryptographic primitives with maximum entropy. In particular, we fabricated arrays of GFETs (Fig. 1) using chemical vapor deposition (CVD) grown graphene. The array contains eight individual GFETs with the same physical dimensions (width and length equal to 1μm) spaced 1μm apart. Arrays were electrically measured and the randomness in the device transfer characteristics were used to generate secure challenge response pairs (CRPs). This was accomplished by first converting the analog current values to a digital 8 bit number. The 8 bit numbers for the devices in an array were strung together to make a 64 bit primitive. Various statistical measures including hamming distance, correlation coefficient, entropy content etc. (Fig. 2) were used to validate the quality of the primitives. Figure 2a shows the entropy of the devices as a function of the gate voltage. Entropy measures the randomness or uncertainty in the CRP. The devices measured have an entropy near unity showcasing a maximum amount of randomness. Figure 2b and 2c are plots of the Hamming distance and correlation coefficient of the CRPs as a function of the gate voltage. These analyses demonstrate the uniformity and uniqueness, respectively. The Hamming distance is a measure of how many bit flips it takes to transform one CRP to another. To ensure maximum uniqueness, Hamming distance should be half the CRP bit length (32 for these primitives). Correlation measures how alike two CRPs are and should ideally be 0. These CRPs have a distribution that peaks close to ideal. As such we can consider the CRPs to be unique, unclonable, and hard to decipher. In addition, we also confirmed the reproducibility of the primitives generated from the same GFET array over time (anti-aging). Finally, we demonstrated how the GFETs can be reconfigured without compromising their entropic strength |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://assets.engr.psu.edu/ESM/docs/esm-today-abstracts/ESM-Today-2019-Program.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
