PNNL

Abstract

Recently much work has gone into utilizing the properties of defects in carbon nanotubes (CNTs). Devices with a single point defect have been used as single molecule sensors,1,2 and CNT network devices with multiple defects have shown higher sensitivity to certain chemicals compared to pristine devices.3 It is an ongoing challenge to link theoretical and experimental studies of the electronic properties of CNT defects. Previous scanning tunneling microscope spectroscopy experiments have shown that defects can locally change the density of states of a CNT.1,4,5,6,7 Electron transport experiments on metallic CNTs with single defects have shown strongly suppressed conductance for electron doping, but much less suppression for hole doping.1 Theoretical work has focused on band structure calculations of CNTs with defects, and often predicts the formation of localized impurity states in and around the band gap.8,9,10,11,12,13,14 A small number of theoretical studies have attempted to link defect structure to electron transport characteristics.8,15 Oxidative defects in CNTs are typically introduced using gas phase or liquid phase treatments such as exposure to heat,16 ozone,17,18 acids,1,19,20,21 or peroxides.22,23 Depending on the oxidant and/or process, a variety of oxidative defects can be created in the sidewall of a CNT such as ethers/epoxides (C-O-C), alcohols (C-OH), ketones/aldehydes (C=O), and carboxylic acids (COOH).24,25 Bulk chemistry using the above methods yields mixtures of these different oxidation states. To generate individual point defects for single molecule sensing, feedback-controlled electrochemical attack1 and local anodic oxidation utilizing voltage pulses from an AFM4,26 or STM27 probe have been applied. None of these techniques offer control over the chemical nature of the defect, and only carboxylic acid groups have been previously identified at the single-defect level.