Selective Detection of Nitroexplosives Using Molecular Recognition within Self-Assembled Plasmonic Nanojunctions

[Image: see text] We demonstrate that the reproducibility of sensors for nitroaromatics based on surface-enhanced Raman spectroscopy (SERS) can be significantly improved via a hierarchical aqueous self-assembly approach mediated by the multifunctional macrocyclic molecule cucurbit[7]uril (CB[7]). Our approach is enabled by the novel host–guest complexation between CB[7] and an explosive marker 2,4-dinitrotoluene (DNT). Binding studies are performed using experimental and computation techniques to quantify key binding parameters for the first time. This supramolecular complexation allows DNT to be positioned in close proximity to the plasmonic hotspots within aggregates of CB[7] and gold nanoparticles, resulting in significant SERS signals with a detection limit of ∼1 μM. The supramolecular ensemble is selective against a structurally similar nitroaromatics owing to the molecular-recognition nature of the complexation as well as tolerant against the presence of model organic contaminants that bind strongly to the SERS substrates.