Optoelectronics with Carbon Nanotubes

Figure III-1. Semi-log plot of drain current as a function of gate voltage at drainsource
biases of -1 V, -3 V, -5 V and -7 V. The dotted arrows indicate the gate sweep
direction.
In order to obtain sufficient signal-to-noise ratio, most of our light emission data is taken
at a high source-drain bias, so the electrical transport is typically near or in the saturation regime,
in the order of 1 to 10 μA. Taking measurements at saturation also makes the data more stable
and reproducible because the results are beyond the influence of charge traps in the substrate. A
typical saturation behavior of our devices is shown as a function of VDS for different VGS values
in Figure III-2. Recall that for metallic SWNTs on a substrate, the saturation current is about 25
μA, while the saturation limit for semiconducting SWNTs depends on the diameter and work
function of metal contacts 56 (note: large-diameter tubes have been shown to be similar to
metallic tubes 55, 57 ). Since we chose small-diameter (i.e., large Schottky-barrier)
semiconducting SWNTs for this study, the saturation current varies significantly from device to
device. Furthermore, since most of our devices are 1-2 μm long, acoustic phonon scattering
(mfp ~ 300-700 nm) and even defects could further reduce the saturation limit.
40