T.V.Afanasieva, I.F.Koval, N.G.Nakhodkin, M.M.Storchak. Proceedings of the second international young scientists' conference on applied physics. June, 19-20, 2002,Kyiv, Ukraine. P.93-94.

# Rotation of Bi-Bi addimer on Si(001)2x1 surface

T.V.Afanasieva, I.F.Koval, N.G.Nakhodkin, M.M.Storchak

*Department of Radiophysics,National Taras Shevchenko University of Kyiv,Volodymyrska 64, Kyiv 03033,Ukraine*

e-mail: afanasieva@univ.kiev.ua

Rotation of Bi dimer on top of dimer row of Si(100)2x1 surface was investigating using ab-initio calculations. Two stable A and B configurations for Bi dimer on top of dimer row were considered. B configuration of Bi dimer is found to be the most stable one while A configuration differs by 0.34 eV from B. Optimal pathway and energy barrier for Bi dimer rotation were found.

## Introduction

Usage of Si/Ge heterostructures is one of the promising approaches in microelecronics. Using surfactants we can dramatically improve quality of Si/Ge interface. Elements of the V group can be used as surfactants [1]. That is why studying interaction of Bi with silicon surface is of a great practical importance. One of basic diffusion acts is rotation of dimer. Two stable configurations are reported in the experiments - A and B [2]. So they are considered. Also the optimal pathway of transformation is studied.

## Methodology

Ab-initio calculation was carried out using restricted Hartree-Fock method. Main calculation program was GAMESS [3]. SBK basis set [4] with SBK effective core potential was applied to Si, Bi and H atom. The silicon surface was modeled by Si29H28 cluster. All bond lengths Si-Si were taken 2.35 A, as the same in the bulk. H atoms were used to saturate dangling bonds on the cluster boundary. Bond lengths Si-H were taken 1.48 A. Coordinates of two top layers of Si and Bi atoms were optimized.

*Fig 1. a) Model of Bi dimer rotation and b) surface of adsorption energy (white arrows indicate optimal pathway for Bi dimer rotation, S is saddle point on this pathway, M is intermediate point)*

## Results

Two stable A and B configurations for Bi dimer on top of dimer row were considered. B configuration of Bi dimer is found to be the most stable one while A configuration differs by 0.34 eV from B. This result is comparable with experimental one (0.17 eV) [2]. The bond length of Bi dimer in B configuration was found to be 3.05 A. Obtained bond length is in excellent agreement with [5,6].

To determine the optimal pathway for the rotation of on-top Bi dimer on the Si(100) 2x1 surface the potential energy surface (PES) was built. PES was mapped out by calculating adsorption energy depending on the angle a, b (fig.1a). Positions of the Bi atoms form an equidistant grid with the spacing between the grid points of 10°. PES for the rotation of the Bi-Bi dimer is shown in fig.1b. The optimal pathway corresponds to consequent motion of ad-atoms. At first one Bi atom moves while another remains at its starting position. Then the first Bi atom occupies its final position (M) and the other starts to move. It is reported that IV group dimers (Si-Ge, Si-Si, Ge-Ge) on Si(001) surface behave in the same way [7]. The determined energy barrier (point S in fig. 1b) for the rotation from B to A is 2.14 eV and the barrier for the reverse direction is about 1.8 eV. This value is overestimated as compared with [2].

To detect other pathways (with Si-Bi exchange) the energy of mixed Si-Bi dimer on the top of Si-Bi/Si-Si structure was estimated. It appeared to be 0.5 eV more, than for Bi-Bi dimer on Si-Si/Si-Si structure in the same position. This energy is less than barrier for rotation.

## Conclusions

Rotation of Bi ad-dimer on Si(100) 2x1 surface was examined. Difference of energy between A and B dimers of 0.34 eV was determined. The optimal pathways for A->B and reversed transformations were defined and the barriers for such rotations were found. Barriers are 2.14 eV for B->A and 1.80 eV for A->B transformations. Possibility of exchange-based transformation was studied.

## Acknowledgements

All calculations have been carried out at the Computer Center of Kyiv National Taras Shevchenko University. Authors thank the Computer Center for the possibility to use the computer cluster system, which has been granted by Intel Corp.

1.Sakamoto K., Kyoya K., Miki K., Matsuhata H. Jpn.J.Appl.Phys. 1993. 32. P.L204.

2.Булавенко С.Ю., Коваль I.П., Мельник П.В., Находкiн М.Г., Симоненко О.М. Сборник докладов 12-го международного симпозиума <Тонкие пленки в электронике>. Харьков (Украина). 2001. С. 303.

3.Schmidt M.W., Baldridge K.K., Boatz J.A., et al. J. Comp. Chem. 1993. Vol. 14. P.1347.

4.Stevens W. J., Basch H., Krauss M. Chem. Phys. 1984. Vol. 81. P. 6026.

5.Franklin G.E., Tang S., Woicik J.C., Bedzyk M.J., et al. Phys. Rev. B. 1995. Vol.52. P. 5515.

6.Gay S.C.A., Jenkins S.J., Srivastava G.P. Surf. Sci. 1997. Vol.384. P.L844.

7.Z.Lu, F.Lui, C.Wang, X.Qin, B.Swartzentruber, et al. Phys. Rev. Lett. 2000. Vol.85, P. 5603.

## Top view | ## Side view |

| |