Skip to main content Accessibility help
×
Home

Peculiar Doping Behavior of Si:Be.

  • Eugen Tarnow (a1), S.B. Zhang (a1), K.J. Chang (a2) and D.J. Chadi (a1)

Abstract

The total energies and structures of a number of Be-induced defects in Si are investigated using ab-initio local density calculations. Our primary results are: 1) The geometry of the isoelectronic center is found to correspond to a [111] substitutionalinterstitial pair (SIP); 2) The low energy defect spectrum includes large Be complexes containing at least one substitutional atom; and 3) Simple bonding rules exist for the stability of the different types of bonds in the material. Thus the Si-Be bond is found to be stable for all defect configurations while the Be-Be bond is metastable.

Copyright

References

Hide All
1. This measurement is attributed to Baker, John E. in reference 2.
2. Crouch, Roger K., Robertson, James B. and Glimer, T.E. Jr, Physical Review B, 5, 3111 (1971).
3. Labrie, D., Timusk, T. and Thewalt, M.L., Phys. Rev. Lett. 52, 81 (1984).
4. Henry, M.O., Lightowlers, E.C., Killoran, N., Dunstan, D.J. and Cavenett, B.C., J. Phys. C: Solid State Physics, 14, L255 (1981).
5. Thewalt, M.L.W., Watkins, S.P., Ziemelis, U.O., Lightowlers, E.C. and Henry, M.O., Solid State Comm. 44, 573 (1982).
6. Killoran, N., Dunstan, D.J., Henry, M.O., Lightowlers, E.C. and Cavenett, B.C., J.Phys C: Solid State Phys, 15, 6067 (1982).
7. Henry, M.O., Moloney, Keith A., Treacy, J., Mulligan, F.J. and Lightowlers, E.C., J. Phys. C: Solid State Phys. 17, 6245 (1984).
8. Davies, Gordon and Henry, Martin, private communication.
9. Robertson, J.B. and Franks, R.K., Solid State Communications, 6, 825 (1968).
10. Kleverman, M. and Grimmeiss, H.G., Semicond. Sci. Technol. 1, 45 (1986).
11. Froyen, Sverre and Zunger, Alex, Phys. Rev. B 34, 7451 (1986).
12. Tarnow, Eugen, Zhang, S.B., Chang, K.J. and Chadi, D.J., Phys. Rev. B, to be published.
13. The formation energies are referenced to c-Si and metallic Be. In particular, the energy we quote for n-atom defects is the energy for n Be atoms to convert from Be metal to the specific configuration with any affected Si atoms going to a Si surface:Eform = Eform[SipBen] - pEform[c-Si] - nEform[Bemetal]
14. Davies, Gordon, J. Phys. C. 17, 6331 (1984).
15. The local mode frequencies for the isolated defects were obtained by assuming that the surrounding Si atoms do not move appreciably. This is justifyable since the Be atom is lighter than the Si atom by a factor of 3. The Be atoms then moved in a tetrahedrally symmetric potential and the frequencies reported are associated with the lowest eigenmode. We performed a test calculation for the substitutional defect allowing one of the Si atoms to move. We found that this changed the local mode frequency by 20 cm-1.
16. Again we approximated the surrounding lattice to be stationary. This should be a good approximation for the antisymmetric state (the effective mass is low), but worse for the symmetric state which has a high effective mass.
17. These estimates were calculated as follows: The upper bound of 0.8 eV was calculated assuming that the atomic coordinates in the configurations in between the local minima were simple linear interpolations. Thus the Si atoms were frozen and the Be atom was not allowed its most favorable path. Then we recalculated the energy barrier by allowing full relaxations of all the surrounding Si atoms. This “adiabatic” energy was found to be 0.4 eV and is too low, because the movement of the Be atom is not a fully adiabatic process.
18. van Kooten, J.J., Weller, G.A., and Ammerlaan, C.A.J., Phys. Rev. B 30, 4564 (1984);
van Kooten, JJ., Ph.D. Thesis, Amsterdam (1987)
and Chantre, A. and Bois, D., Phys. Rev. B31, 7979 (1985).
19. For a theoretical treatment see Scheffler, Matthias, Festkörperprobleme 29, 231 (1989).

Metrics

Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed