Robustness of logic gates and reconfigurability of neuromorphic switching networks

Nanoparticle networks with functional molecular links that show current-voltage characteristics (IVC) with negative differential resistance (NDR) can be trained to perform XOR-AND logic gates (Husband et al. [1]; Skoldberg and Wendin [2]). In this work we investigate the robustness of the Nanocell network by removing links until desired logic gates no longer can be configured or operated within our simulation of the network. We present results for the robustness of XOR-AND configured (halfadder) Nanocells, as well as the effects of varying the IVC and NDR characteristics of the linker molecules.