Quantum Simulation of Amorphous Silicon: Preparation, Structure and Properties
- PDF / 350,237 Bytes
- 6 Pages / 420.48 x 639 pts Page_size
- 55 Downloads / 183 Views
QUANTUM SIMULATION OF AMORPHOUS SILICON: PREPARATION, STRUCTURE AND PROPERTIES
L. COLOMBO and G. SERVALLI UniversitA di Milano, Dipartimento di Fisica, via Celoria 16, 1-20133 Italy
Milano,
ABSTRACT Within a tight-binding molecular dynamics scheme we investigate pure amorphous silicon (a-Si) obtained by direct quenching from the melt. Using different rates for the cooling process, we demonstrate that both structural and electronic properties of a-Si depend on the sample preparation. Possible size-effects are also investigated using 64- and 216-atom supercells. Finally, we discuss the reliability and transferability of the present scheme for large scale simulations of covalent materials.
INTRODUCTION Several basic problems about the physical properties of amorphous silicon (a-Si) still remain open, despite a detailed microscopic characterization of a-Si has been very recently obtained using ab-initio molecular dynamics (MD).[1] The main goals of the present paper are: (i) to understand the dependence of the structural properties of a-Si upon the sample preparation; (ii) to give a detailed characterization of the chemical bond in the amorphous network; (iii) to describe the possible interplay between the macroscopic disorder of the host structure and short-range-order (SRO) features. The above questions are addressed performing tight-binding molecular dynamics (TBMD) simulations.[2,3] Thanks to the reduced computational cost of the TBMD approach, we investigate different cooling rates for the liquid-tosolid transition involving comparatively long simulation times (of the order of 102 picoseconds). Moreover, by using supercells (SC) containing up to 216 atoms, we study the convergence of the results against the size of the sample. Finally, contrary to existing speculations [1], we demonstrate that the TBMD approach, despite it is a semi-empirical method, is a reliable, accurate and transferable scheme for quantum simulations of covalent semiconductor materials in a wide range of temperatures and for a large variety of structures: it turns out to be one possible tool to bridge the gap between classical empirical simulations and quantum mechanical first-principles ones.
COMPUTATIONAL FRAMEWORK In the TBMD scheme the total potential energy of a N-particle system, Etot, is given by two terms [4]: (i) a band-structure contribution Ebs and (ii) a sum of short-range nearest-neighbour two-body effective potentials V(r•): Etaz(rl,r2,....rs) = Eba(ri,r2,....rs) +
.j•iV(r•u) (1)
Mat. Res. Soc. Symp. Proc. Vol. 291. 01993 Materials Research Society
62
where ri (i=1,...,N) are the atomic positions, rij=Iri-rjl are the relative interatomic distances and Ebs is the sum over all the single-particle energies of the occupied states En(k) Ebs = 2
-kn cc".En(k)
(2) obtained from a semi-empirical tight-binding (TB) Hamiltonian (in eq.(2) the sum over the k vectors equals the integration over the Brillouin zone of the MD cell). In the present work we adopted the Goodwin et a]. approach [51, where both V(rij) and the scaling law
Data Loading...
