FLORO RESEARCH GROUP

Figure 1

(click to enlarge)



Figure 2

(click to enlarge)

Thermal and Electronic Transport Processes in
Monolayer-scale Chemically Ordered Semiconductor Films


Generously supported by National Science Foundation

PI: Pamela Norris, Co-PI: Jerry Floro
Graduate Students: Jatin Amayta (with Floro), Mackenzie Sinden-Redding and Justin Smoyer (with Norris)
Project Status: ongoing

Goals: In this work, we will investigate how atomic-scale chemical ordering in Si-Ge alloys affects thermal and electronic transport relevant to thermoelectric materials for energy harvesting.

Technical Approach: Si-Ge alloys are known to undergo a metastable chemical ordering process during molecular beam epitaxy (MBE) on Si (001). This ordering, which leads to an L11-like structure, results from atomic processes occurring at the 2x1 reconstructed growth surface. In metals, ordering is known to reduce alloy scattering of electrons, thereby enhancing electrical conductivity. Although this would seemingly also enhance the thermal conductivity, which is undesirable for thermoelectrics, our MD calculations suggest that at high temperatures, thermal conductivity is close to the alloy limit due to increased Umklapp scattering associated with the larger Brillouin zone in the ordered phase, see Figure 1 and “Reducing thermal conductivity of binary alloys below the alloy limit via chemical ordering”, J. C. Duda, et al., J. Phys. Conden. Matter, 23, 205401, 2011.

Key Results: We have grown Ge0.5Si0.5 films on Si (001), exploring a range of growth temperatures, deposition rates and film thicknesses. Chemical ordering is observed in x-ray diffraction via the presence of the ½(111) superlattice reflection. Despite the wide range of parameters investigated here, the order parameter never exceeds 0.07 (this is the volume-averaged order parameter for the entire film). We recently have grown similar films on Ge (001) substrates, where the tensile mismatch strain of the film suppresses quantum dot formation, reducing RMS roughness by more than an order of magnitude (Fig. 1). We have at least one instance where the order parameter reaches 0.21 - still low, but a significant leap from the prior limitation on Si substrates. We are currently trying to understand how the ordering-variant domains are distributed throughout the film using high-resolution transmission electron microscopy and scanning transmission electron microscopy (Fig. 2).