Theoretical investigation of silicon nanowires: Methodology, geometry, surface modification, and electrical conductivity using a multiscale approach

The structural and electronic properties of hydrogenated silicon nanowires (SiNWs) oriented in < 100 >, < 110 >, < 111 >, and < 112 > directions are investigated systematically using a multiscale approach: geometry optimization is done by a semiempirical method and electronic band structure is calculated by density-functional theory with Gaussian basis set. The calculated band gaps agree very well with the available experimental data. We propose that the multiscale approach is an accurate and effective way for calculating the structural and electronic properties of SiNWs with diameter up to 3 nm. Besides, we found that surface modification of the SiNWs using the hydroxyl and fluoro groups can strongly reduce the band gap by as much as 1 eV, and more interestingly, alter the gap nature. On the contrary, the < 110 > SiNWs exhibiting different patterns at the cross section do not show significant difference in their band gaps (< 0.09 eV), indicating that the electronic structures of SiNWs are much more sensitive to surface modification than the change of cross section. Moreover, SiNWs of different orientations exhibit different degrees of band gap reduction upon surface modification, in which the < 110 > SiNW demonstrates the highest sensitivity. The result indicates that SiNWs oriented in < 110 > direction are the better candidate for sensor application. On the contrary, the < 111 > SiNW is found to be very tight to structural change with diameter and the most reluctant to surface modification, showing that it is structurally stable and rather inert. In addition, from the I-V curves of the SiNWs with surfaces modified with hydroxyl groups, which are calculated by the nonequilibrium Green's function theory, we found that the electrical conductivity of the SiNWs is highly chemical sensitive. Besides, the phenomenon of negative differential resistance is observed in the I-V curves.