The thermal conductivity of nanostructured materials can differ from that of their bulk single-crystal counterparts by orders of magnitude, with major impacts on diverse applications from transistors to energy conversion. Here I will describe two examples:

(1) Although suspended graphene has been reported to have very high in-plane thermal conductivity, most applications would require graphene to be supported or encased within dielectric layers. Our measurements of encased graphene and ultrathin graphite show that the constraints of the encasing layers reduce the in-plane thermal conductivity by at least a factor of 10 as compared to bulk graphite.

(2) Bulk nanocomposites are promising thermoelectric materials because they can potentially combine the performance advantages of nanostructuring with scalable, low-cost synthesis. To clearly quantify the effects of grain boundaries in reducing the phonon thermal conductivity in such materials, we measured undoped nanocrystalline silicon as a model system. The results show that the effective boundary scattering length is somewhat smaller than the average grain size, and reveal a previously unidentified frequency dependence which we show is consistent with asymptotic analysis of atomistic simulations from the literature.