Substitutional group III and group V elements, though commonly used as shallow dopants in bulk silicon, have a limited efficiency in silicon nanocrystals. In this work, we use firstprinciples models of 1.5 nm nanocrystals with hydride-and silanolterminated surfaces to understand how oxidation influences the segregation and deactivation of dopants at the surface and the dopant binding energies. We show that the surface oxygen layer changes drastically the radial dependence of the dopant formation energy both for donors and for acceptors, but that, independently from the oxidation, dopant diffusion does not take place at operating conditions. Additionally, we show that the oxidation increases the electron binding energy of the P, As, and Sb and decreases the hole binding energy of B, Al, Ga, and In.
■ INTRODUCTIONThe ability to control the type and concentration of charge carriers in silicon by doping with extrinsic impurities has been central to its success in electronics. It has recently been demonstrated that silicon nanocrystals (NCs) can also be doped n-type or p-type with phosphorus and boron, respectively. 1−3 Growth of doped nanocrystals by nonthermal plasma synthesis can be achieved by addition of phosphine and diborane to the precursor gas mixture, and incorporation efficiencies close to 100% have been reported.Still, the doping of nanocrystals is hindered by some difficulties not present in their bulk counterparts. One of them is the higher dopant activation energy. If the nanocrystal radius is smaller than the Bohr radius of the electron or hole in bulk silicon, the excess electron or hole is considered to be in the strong quantum confinement regime. 4 Due to the quantum confinement, the binding energy of the electron or hole to the ionized dopant can be more than 1 order of magnitude larger than in bulk. 5,6 The smaller the nanocrystal the smaller the effective radius a e characterizing the exponential decay of the hydrogen-like wave function of the bound electron or hole s state, ψ(r) ∼ exp(−r/a e ), where r is the distance between the two. 5 In medium-sized nanocrystals (with ψ(r) ∼ 0 at the surface), even though the quantum confinement is weak, the screening by image charges at the surface of the nanocrystal or at the interface between the nanocrystal and the outer oxide shell may also contribute to increase the dopant activation energy. 7,8 Another problem is to guaranty that dopant incorporation occurs at substitutional positions in the nanocrystal core, where the dopant becomes active. Depending on the nanocrystal size and specific growth method, 80−95% of the P atoms can stay on the nanocrystal surface. 2,3 If the nanocrystals are covered by an oxide layer, as much as 95% can be located there, presumably because P atoms segregate to the surface during growth, becoming incorporated in the oxide during the subsequent surface oxidation process. 1 Once at the surface, P (and B) may assume three-fold coordination and become electrically inactive. 9,10 In this way, surface dangling bonds can be passiv...