Computational design of gas phase explosive sensors

Graphene based sensors have great potential in the trace detection of hazardous gases. The application of graphene based sensors in the trace detection of explosives has seen relatively limited study, due in part to the difficulties of conducting experiments using the nitramine and aromatic explosives of central interest. Computational studies of explosive sensors are not subject to hazardous materials handling constraints, and may be used to complement experimental research on the development of low weight, low power, graphene-based sensors. In this dissertation, the sensing properties of graphene nanoribbons (GNRs) and its derivatives have been investigated for three widely used explosives and four background gases, using ab initio computational models. A variety of different design considerations have been studied, including the effects of doping, nanopatterning, mechanical deformation and junctions on sensor performance. Based on the simulation results, an innovative and compact dual-mode sensor configuration is proposed, which may significantly enhance chemiresistive sensor selectivity and significantly reduce the sensor count in sensor array systems. In addition, an improved ab initio molecular dynamics method has been developed, by applying nonholonomic Hamiltonian methods to model a mixed classical-quantum system. This adaptive method allows for the thermodynamically consistent modeling of complex sensor physics, including multi-scale coupling, multi-energy domain coupling, and deforming control volumes, all issues of considerable interest in sensor design.