Fig. 4: Evaluation of the proposed workflow. | Nature Communications

Fig. 4: Evaluation of the proposed workflow.

From: Demonstration of high-reconfigurability and low-power strong physical unclonable function empowered by FeFET cycle-to-cycle variation and charge-domain computing

Fig. 4

a The Gaussian-like distribution of \({V}_{{{{\rm{X}}}}}\) difference in two cycles (\(\Delta {V}_{{{{\rm{X}}}}}\)) ensures uniform randomness of the PUF cell states. b \(\Delta {V}_{{{{\rm{X}}}}}\) map and cell states before and after rewrite. c After rewrite, the distinct \({V}_{{{{\rm{X}}}}}\) separation between PUF cells with different outputs enhances repeatability. d \({V}_{{{{\rm{R}}}}}\) is set to 1.2 V to maximize \(\Delta {V}_{{{{\rm{X}}}}}\) for C2C variation extraction. e \({V}_{{{{\rm{X}}}}}\) is the XNOR result of the input challenge (\(C\)) and the cell state (\(S\)), where the on/off ratio is maximized at 1.2 V read voltage. f \({V}_{{{{\rm{SL}}}}}\) shows high linearity and can be sensed with an appropriate \({V}_{{{{\rm{ref}}}}}\). \({V}_{{{{\rm{DD}}}}}\) is set to 0.5 V to save energy and provide a sufficient sense margin. Source data are provided as a Source Data file.

Back to article page