As materials are confined to low dimensions with a size comparable to the scattering mean free paths, the thermal conductivity is often reduced due to increased boundary scattering. The reduced thermal conductivity is desirable in some applications such as thermoelectric cooling, but is often unwanted for others especially for nanoelectronic devices. An exception of this scaling trend is carbon nanotubes (CNTs). Due to the unique crystalline structure, boundary scattering is nearly absent in CNTs, giving rise to super high thermal and electrical conductivity that makes the CNT an ideal candidate for replacing Cu in future VLSI interconnects. The potential electronic applications have inspired several groups to employ a variety of techniques for measuring the Seebeck coefficient , specific heat , and thermal conductivity  of CNT bundles and mats. Estimated thermal conductivity from these measurements is significantly reduced by numerous tube-tube junctions in the sample and is much lower than the theoretical expectation .