Designing Real Nanotube-Based Gas Sensors

Using a combination of density functional theory and recursive Green‘s functions techniques, we present a full description of a large scale sensor, accounting for disorder and different coverages. Here, we use this method to demonstrate the functionality of nitrogen-rich carbon nanotubes as ammonia sensors as an example. We show how the molecules one wishes to detect bind to the most relevant defects on the nanotube, describe how these interactions lead to changes in the electronic transport properties of each isolated defect, and demonstrate that there are significative resistance changes even in the presence of disorder, elucidating how a realistic nanosensor works.

Figure 1

Different ${\mathrm{CN}}_x$ CNTs after relaxation: (a) 3NV, (b) 4ND in the cases 1) without and 2) with ammonia. For clarity the structures have been rotated around the tube axis and only the central section of the supercell is shown. c) Example of a disordered ${\mathrm{CN}}_x$ structure. Color code: carbon: black, nitrogen: blue, hydrogen: light blue.

Figure 2

Transmission coefficients (a), (b) and DOS (c), (d) for ${\mathrm{CN}}_x$ nanotubes with three- and four-nitrogen defects. The solid (dashed) line shows the results without (with) ammonia dissociated on the defect.

Figure 3

Average of the conductance (in units of ${\mathrm{G}}_0=2e^2/h$), as a function of ammonia concentration for 3NV in ${\mathrm{CN}}_x$ nanotubes. The inset shows a histogram of conductance values for the 3NV defects in the presence of ammonia.

Figure 4

Average of the conductance (in units of ${\mathrm{G}}_0=2e^2/h$) as a function of ammonia concentration for 4ND. The inset shows typical transmission curves as a function of energy with 0 % (solid line) and 100 % (dashed line) coverage.