SRC Quantum Lunch
Probing local, strain-induced shifts in the Si conduction band on the nanoscale with X-PEEM
This project is focused on locally strain engineering silicon nanomembranes. By “nanomembranes” we mean single crystal sheets of Si that are very thin, only a few 10s to 100s of nanometers thick. When Si is made this thin, it becomes flexible (stretchable and bendable) while maintaining it's crystalline properties (e.g. elastic anisotropy, band structure, etc). For example, making a sandwich of Si/SiGe/Si, where the SiGe alloy layer is deposited in a compressively strained form, will result in a strain of the outer Si layers when the membrane is released from its original substrate. Previous work at the SRC used x-ray absorption spectroscopy (XAS) on the high energy resolution VLS-PGM beamline to measure strain-induced conduction band shifts in such membranes .
We use the SHINX at the SRC to similarly probe the conduction band of silicon with the spatial resolution afforded by the microscope. We perform similar XAS measurements, taking a “movie” of our sample as a function of incident beam energy and collecting a complete spectrum for each pixel in the image. Where in the previous work we used uniformly deposited films of SiGe alloy to strain our Si membranes, here we deposit localized Ge stressors. Ge forms small (30-100 nm) “huts” when grown on Si because strain energy builds up so quickly in the Ge that it cannot form smooth layers. This is somewhat akin to the way water beads up on a non-stick frying pan. When these huts are grown on free-standing Si nanomembranes, they locally strain the Si in the region beneath and around the huts. An interaction between the hut's strain field and the elastic anisotropy of the membrane even leads to an interesting ordering phenomenon . This local strained region will also have a locally modified electronic bandstructure, which is what we are looking at with the SPHINX.
One potential use for such a structure would be as a thermoelectric material. Thermoelectric devices convert a thermal gradient into a voltage, (i.e. heat into electricity with no moving parts) or vice versa (i.e. refrigeration with no moving parts). Nanomaterials are very interesting thermoelectric materials because nanoscale structures generally have very low thermal conductivity. Taking a nanoscale object like a Si nanoribbon and adding a periodic band structure modulation such as what we expect to be casued by the Ge huts can create an electronic superlattice. Engineering an electronic superlattice in a nanowire could give us a way to tune the electronic density of states to improve thermoelectric properties of the wires. We discuss the sizes of huts and the amount of strain needed to create such superlattices in Huang et al . We use the PEEM to examine how much strain-induced conduction band shift is present in our samples and also to look at membranes that have been locally strained in other ways like simple mechanical bending and stretching.
 M.M. Roberts et al. Nature Materials *5*, 388 (2006).
 Euaruksakul et al. Physical Review Letters. *101,* 147403 (2008).
Kim-Lee et al. Physical Review Letters. *102*, 226103 (2009).
 Huang et al. ACS Nano. *3*, 721 (3009).