Thermal and electrical transport in a new class of nanocomposites composed of random isotropic two-dimensional ensembles of nanotubes or nanowires in a substrate (host matrix) is considered for use in the channel region of thin-film transistors (TFTs). The random ensemble of nanotubes is generated numerically and each nanotube is discretized using a finite volume scheme. To simulate transport in composites, the network is embedded in a background substrate mesh, which is also discretized using a finite volume scheme. Energy and charge exchange between nanotubes at the points of contact and between the network and the substrate are accounted for. A variety of test problems are computed for both network transport in the absence of a substrate, as well as for determination of lateral thermal and electrical conductivity in composites. For nanotube networks in the absence of a substrate, the conductance exponent relating the network conductance to the channel length is computed and found to match experimental electrical measurements. The effective thermal conductivity of a nanotube network embedded in a thin substrate is computed for a range of substrate-to-tube conductivity ratios. It is observed that the effective thermal conductivity of the composite saturates to a sizeindependent value for large enough samples, establishing the limits beyond which bulk behavior obtains. The effective electrical conductivity of carbon nanotube-organic thin films used in organic TFTs is computed and is observed to be in good agreement with the experimental results. DOI: 10.1115/1.2709969


nanotube, thin film transistor, nanocomposite, percolation

Date of this Version

April 2007