Ab-initio calculations for a realistic sensor: A study of CO sensors based on nitrogen-rich carbon nanotubes

A M Souza,A Fazzio,A R Rocha,A J R Da Silva

doi:10.1063/1.4739280

Abstract

The use of nanoscale low-dimensional systems could boost the sensitivity of gas sensors. In this work we simulate a nanoscopic sensor based on carbon nanotubes with a large number of binding sites using ab initio density functional electronic structure calculations coupled to the Non-Equilibrium Green's Function formalism. We present a recipe where the adsorption process is studied followed by conductance calculations of a single defect system and of more realistic disordered system considering different coverages of molecules as one would expect experimentally. We found that the sensitivity of the disordered system is enhanced by a factor of 5 when compared to the single defect one. Finally, our results from the atomistic electronic transport are used as input to a simple model that connects them to experimental parameters such as temperature and partial gas pressure, providing a procedure for simulating a realistic nanoscopic gas sensor. Using this methodology we show that nitrogen-rich carbon nanotubes could work at room temperature with extremely high sensitivity.

Highlights

The possibility of detecting very low concentrations of chemical species is a fundamental issue for a variety of processes such as in, industrial and environmental monitoring, and medicine.[1,2,3] In particular, gas sensors are desired for detecting gas molecules which could present high toxicity such as carbon monoxide
We have presented a methodology whereby one is able to use ab initio atomistic methods combined with Non-Equilibrium Green’s Function electronic transport calculation to realistically simulate a nanoscopic gas sensor. This is achieved by i) first determining the binding properties of the molecule one wishes to detect via density functional theory (DFT) calculations, ii) calculating the electronic transport properties of an isolated defect and, most importantly, of a disordered system where the binding sites are randomly distributed along the device, which in turn allows for different gas coverages to be taken into account. iii) we connect our conductance results to experimental parameters such as temperature and gas pressure
As a point in case we have determined that the 4ND defect in CNx nanotubes is a very promising system to be used as gas sensor to detect carbon monoxide

Summary

INTRODUCTION

The possibility of detecting very low concentrations of chemical species is a fundamental issue for a variety of processes such as in, industrial and environmental monitoring, and medicine.[1,2,3] In particular, gas sensors are desired for detecting gas molecules which could present high toxicity such as carbon monoxide. Solid-state sensors operate by resistance changes due to alien species binding to the surface of the device From this perspective it is important to have solid state devices which could operate at room temperature and with higher sensitivity than the present semiconductorbased ones.[4,5] Nanoscopic one-dimensional structures could provide such progress. Due to their high surface area to volume ratio the adsorption of molecules on the surface of one-dimensional systems - even in small concentrations - can lead to significant changes on their electronic and transport properties. Experimental[20] as well as theoretical[21] works have demonstrated that those nitrogen-rich nanotubes could be assembled into devices that detect a variety of gases

THEORETICAL FRAMEWORK

Energetic properties and absorption process study

Transport properties of a single defect

Transport properties of the disordered system

Room temperature detection

CONCLUSION