A main issue in information technology affecting heterogeneous integration and autonomous systems is the fact that they need advanced thermal management and, in the autonomous case, their own energy source (energy harvesters). In general, in energy conservation and generation and in efficient energy transformation, among others, the control of heat dissipation is of paramount importance. From data centres to basic transistors with a sub 20 nm gate width, the issue is thermal management involving heat transport in the plane and across interfaces, all of which is exacerbated as dimensions become smaller.
Our work addresses the physics underpinning heat transport by focusing on the changes in the dispersion relations of phonons in freestanding membranes down to 8 nm thickness. We have observed and simulated the strong modifications of the dispersion of phonons in model structures, namely freestanding ultra-thin crystalline silicon membranes, which happen to be the materials at the centre of information and communication technologies.
Our results show that to simulate, and eventually design, new components and circuits, thermal conductance/conductivity and other properties affected by the density of states, including the spectral and geometric dependencies of phonon properties, must be taken into account. Although our studies are only on Si, they are equally applicable to other thin film materials and nanostructures and even more so to multilayer structures, since the acoustic impedance and contrast changes with type, shape, geometry and thickness of each layer.
This is probably the first report of Lamb waves in the sub-THz regime, which will be of interest to the acoustics community and some sectors of mechanical engineering.
Our results impact directly not only heat dissipation issues in nanoelectronics but also the physics of nano-scale thermoelectricity. Both are phonon-dependent and naturally affected by temperature and frequency.
All sensors using nanostructures which require low power electronics and energy harvesting to be really "autonomous", sooner or later, hit upon the issue of thermal management. Our results will help in the modelling and design of nanotechnology-enabled sensors, especially to advance knowledge on lattice vibrations and heat transport in nanosystems.
The simulations, theories and experiments are intricate and involved knowledge from mechanical and electrical engineering, solid state physics, statistical mechanics, inorganic chemistry and even biophysics. For example, artificial

When an organism develops its shape or heals wounds, or when tumours metastasize, cells undergo massive collective movements. Despite decades of research, the mechanisms underpinning these movements remain poorly understood. It has been argued that chemical stimuli and chemical gradients alone cannot explain this form of cellular migration, and some evidence has implicated a role for mechanotransduction -- the perception and reaction to a mechanical force. Using new technologies developed in our own group, we discovered that cells transmit forces from their leading edge, creating a stress wave that propagates through the mass of expanding tissue.

To study collective cell migration in a controlled manner we used soft lithography techniques. Using these techniques we were able to produce polydimethylsiloxane membranes with a rectangular opening that was placed on top of a polyacrylamide gel coated with collagen I. We then grew Madin-Darby Canine Kidney cells to confluence in the rectangular opening and, as the membrane was removed, cells spontaneously migrated. To measure the forces driving such migration we used monolayer stress microscopy, a technique developed in our own group.

Cells at the edges of the epithelial monolayer were the first to migrate and to generate forces. Surprisingly, these forces were transmitted from cell to cell through intercellular junctions in a wave-like manner. The wave velocity was ultraslow, roughly one millimetre per day. As such, these waves are certainly among the slowest mechanical waves ever described.

To study the origin of these mechanical waves we built a mathematical model that captured the observed phenomenology using two purely mechanical assumptions: that a cell acquires a motile phenotype only when an adjacent cell creates space or pulls on the shared inter-cellular junction; and that the cell contains a strain threshold that, once exceeded, results in the cytoskeleton first reinforcing its stiffness and then breaking down (fluidizing). These findings indicate that although chemical stimuli and/or gene expression could be involved in epithelial monolayer migration, these are not necessary for the generation of mechanical waves.

This study brings us one step closer to understanding how cells migrate, and thus a stage nearer to understanding the dynamics of tumour cells and the physical mechanisms they use to break away and metastasize.