D. Mangelinck et al., INFLUENCE OF THE ADDITION OF CO AND NI ON THE FORMATION OF EPITAXIAL SEMICONDUCTING BETA-FESI2 - COMPARISON OF DIFFERENT EVAPORATION METHODS, Journal of applied physics, 83(8), 1998, pp. 4193-4201
beta-FeSi2 films containing 4-25 at. % of Co or Ni have been grown on
(111) Si using three different evaporation methods: deposition of the
metal film at room temperature followed by solid state reaction (solid
phase epitaxy), deposition of the metal on a heated wafer [reactive d
eposition epitaxy: (RDE)] codeposition of metal and Si in the stoichio
metric proportions [molecular beam epitaxy: (MBE)]. The films have bee
n analysed with Rutherford backscattering spectrometry, x-ray diffract
ion, secondary ion mass spectroscopy, and scanning electron microscopy
. We show that beta-FeSi2 can be obtained with the three methods, howe
ver the quality, the epitaxial character and the morphology of the fil
ms differ. By far, the best results are obtained by MBE. In that case
the codeposition of a Fe(Ni) alloy and Si at 550 degrees C leads to a
large, epitaxial, mirror like beta-Fe(Ni)Si-2 layer with a homogeneous
Ni concentration. Detailed analysis of the Fe(Ni) and Fe(Co)/Si react
ion and comparison with the Fe/Si one shows that: (i) Ni and Co do not
modify the temperature of formation and the stability of the differen
t Fe silicides providing the solubility limits in the different phases
are not exceeded (e.g., 6 and 12 at. % in beta-FeSi2), (ii) the first
stages of the reaction are characterized by the formation of a nonhom
ogeneous layer of FeSi. We did not observe the formation of a Si solid
solution in Fe or of the ordered Fe3Si phase, (iii) the transformatio
n from FeSi to beta-FeSi2 induces a degradation of the beta-FeSi2 laye
r morphology. This explains why the processes (RDE and MBE) bypassing
the formation of FeSi give better epitaxial layers. (C) 1998 American
Institute of Physics.