Recent advances in nanotechnology have yielded materials and structures that offer great potential for improving the sensitivity, selectivity, stability, and speed of next-generation chemical gas sensors. To fabricate practical devices, the "bottom-up" approach of producing nanoscale sensing elements must be integrated with the "top-down" methodology currently dominating microtechnology. To demonstrate this approach, a single-crystal SnO2 nanowire sensing element is mounted onto a microhotplate gas sensor platform. This prototype sensor possesses encouraging performance aspects including reduced operating temperature, reduced power consumption, good stability, and enhanced sensitivity.
Chemical gas sensors that utilize semiconducting metal oxides as transducers have been studied for some time, yet the mechanisms for their operation are not fully understood due to the complexities inherent in the fabrication of polycrystalline films of chemically sensitive metal oxides. The recent introduction of methods for synthesizing single-crystal nanoscale structures, such as nanowires, makes possible direct observations of chemical sensing reactions on simpler, more easily modeled structures. The goals of this applied research project are to integrate such nanostructures with existing chemical sensors based upon MEMS (microelectromechanical systems) technology, demonstrate their function, and describe their mechanisms of operation more precisely than currently feasible with polycrystalline metal oxide film sensors.
These experiments were performed by generating tin oxide nanowires, transferring them to individually addressable microheater sensor platforms, then using FIB (focused ion beam) platinum deposition to attach them to the platinum electrodes. This completed sensor was used to perform a series of conductometric sensing studies to gauge sensitivity and selectivity of the nanowires as sensors. The target analytes were carbon monoxide (tested in a range from 25 μmol/mol to 100 μmol/mol) and ammonia (tested in a range from 100 nmol/mol to 100 μmol/mol) in dry air backgrounds. It was found that the sensitivity to ammonia was much greater than that to carbon monoxide. A further examination of the sensitivity and dynamic detection range for ammonia found that the signal-to-noise ratio allows detection as low as 30 nmol/mol. The conductance response ΔG=Gammonia-Gair follows a power law relationship with the analyte concentration C: ΔG≈Cα, with the exponent approximately equal to 0.5. This relation holds through an analyte concentration excursion spanning four orders of magnitude.
According to a published model (N. Barsan and U. Weimar, J. Electroceram. 7, 143, (2001)), three conclusions can be implied: (1) that the approximation of the conduction channel R≈λD (where R is the radius and λD the Debye length of the material at the given temperature) can be applied to the nanowire, (2) that singly charged O− is the predominant ionosorbed species that reacts with reducing gases, and (3) since signal saturation did not occur, θ(NH3)<<θ(O−) (θ is surface coverage) in the tested analyte concentration range. Since the nanowire is a single-crystal conducting channel, this model can be readily applied. Conventional polycrystalline nanoparticle films, on the other hand, exhibit lesser-defined percolation between multiple grains, complicating the analysis. Due to the well-defined composition, crystallinity, and morphology of single-crystal nanowire sensing elements, the performance of such conductometric sensors can be modeled more simply than polycrystalline counterparts, which is a considerable advantage compared to the phenomenological approach currently used to describe film sensors.
Applied Physics Letters 91, 063118 (2007)