Forschung im HLRN-Verbund 2011

121Figure 1: Cross-section of an undoped, natively oxidized Si nanowire showing that the most probable localizationregions of hole (top of the valence band, TVB) and electron (bottom of the conduction band, BCB) carriersare at the Si/SiO x interface.esting, since the carriers of electronic current, providedby the dopants, face now two difficulties intheir transport path: the disorder in the silicon coreand the presence of an electronegative oxide shell.So, which of these effects will have the most prominenteffect on the conductivity of the strucutures?To answer this question, we have performedlarge-scale quantum mechanical calculations ofthe SiNW transport properties on the parallel supercomputersat the HLRN. While most previousstudies of non-oxidized wires relied on more or lessapproximated approaches, our calculations arebased on Density Functional Theory (DFT), whichis a computationally very expensive approach fortransport calculations even with simple systems.However, we realized that this level of precision isneeded for our work, covering subtle effects suchas dopant and structural backscattering, and relyingon the precise determination of the transmissionchannels for conductance.Our results achieved so far [5] clearly show thatthe conductance of the oxidized wires is decreasedwith respect to the case of ideally crystalline, H-terminated wires, and the main reason for this decreaseis the oxidation-induced disorder. The overallconductance decrease is very similar for B andP dopants, despite the fact that the transmissionof P doped wires is less influenced by structuraldisorder in the absence of an oxide shell. If anoxide shell is present, though, it attracts the extraelectrons carried by the P atoms and changes thetransmission channel from a localized (less sensitiveto disorder) to a delocalized form.Future work will be devoted to understandingthe mechanisms of detection of biomolecules withsuch oxidized SiNWs. This will require furtherlarge-scale quantum transport calculations andthus call for extended use of the HLRN computationalresources.More Information1. W. Lu, C. M. Lieber, “Nanoelectronics from thebottom up”, Nature Mater. 6, 841-850 (2007).2. E. Stern et al., “A nanoelectronic enzyme-linkedimmunosorbent assay for detection of proteinsin physiological solutions”, Small 6, 232-238(2010).3. V. Schmidt, J. V. Wittemann, U. Gösele,“Growth, thermodynamics, and electrical propertiesof silicon nanowires”, Chem. Rev., 110,361-388 (2010).4. L. Colombi Ciacchi, D. J. Cole, M. C. Payne, P.Gumbsch, “Stress-driven oxidation chemistry ofwet silicon surfaces”, J. Phys. Chem. C Letters112, 12077-12080 (2008).5. M. Koleini, L. Colombi Ciacchi, M.-V.Fernández-Serra, “Electronic transport in nativelyoxidized silicon nanowires”, submitted toACS Nano (2011).FundingUniversity of Bremen; German Research Foundation(DFG)Ingenieurwissenschaften