Crossover energetics for halogenated Si(100): Vacancy line defects, dimer vacancy lines, and atom vacancy lines

Abstract

We investigated surface patterning of I-Si͑100͒-͑2 ϫ 1͒ both experimentally and theoretically. Using scanning tunneling microscopy, we first examined I destabilization of Si͑100͒-͑2 ϫ 1͒ at near saturation. Dimer vacancies formed on the terraces at 600 K, and they grew into lines that were perpendicular to the dimer rows, termed vacancy line defects. These patterns were distinctive from those induced by Cl and Br under similar conditions since the latter formed atom and dimer vacancy lines that were parallel to the dimer rows. Using first-principles density functional theory, we determined the steric repulsive interactions associated with iodine and showed how the observed defect patterns were related to these interactions. Concentration-dependent studies showed that the vacancy patterns were sensitive to the I concentration. Dimer and atom vacancy lines that were elongated along the dimer row direction were favored at lower coverage. Atom vacancy lines dominated at ϳ0.8 ML, they coexisted with dimer vacancy lines at ϳ0.6-0.7 ML, and dimer vacancy lines were exclusively observed below ϳ0.5 ML. These surface patterns reflect the mean strength of the adatom repulsive interactions.

Energy scaling and surface patterning of halogen-terminated Si(001) surfaces

“…For large halogens (I and At), we predict a crossover to a new type of one-atom wide vacancy-line defects and regrowth chains. Our predicted VLD for iodine is now confirmed experimentally [7].…”

“…Hence, b % a/2 % 3.2n 2 meV is, practically speaking, as good as direct DFT calculation. Recently our scaling-and DFT-predicted results for iodine have been confirmed by experiment [7].…”

“…Also, a fully-covered silicon surface may roughen only by etching, because free surface sites are needed for other types of surface kinetics (indeed, steric roughening of a fully-covered surface is not observed in experiment [7]). If one consider a halogen-covered Si sample at a finite temperature in vacuum, there is also a loss of halogen atoms from surface over time.…”

“…Surprisingly, this simple conclusion is correct for Cl, Br, and I, and fails for F, which is highly reactive and roughens Si surface by etching. However, other STM observations [7] demonstrated that neither the number of S B steps nor the number of passivating atoms remains constant during experiment. Below we allow the possibility of creating new S B steps, and assume, that the number of passivating atoms might change, but the experimentally assessed energy of S B steps takes this non-conservation into account.…”

Structural database for reducing cost in materials design and complexity of multiscale computations

“…We found, that those interactions increase with the size of halogen atoms: they scale as n 2 , where n is the period of the halogen in Mendeleyev Periodic Table. Using this scaling, we estimated interaction for iodine on Si(001) surface (later we calculated energies of iodine surface patterns from the first principles and confirmed this estimation), compared this interaction with the energies of step defects on silicon surface, and predicted that iodine coverage of Si(001) should result in a new type of linear surface defect: the vacancy line defect [16]. This predicted defect is now observed experimentally [17]. This example illustrates how reuse of previously calculated structural data can lead to new discoveries.…”