Optical Properties of Semiconducting Silicides
- PDF / 408,178 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 100 Downloads / 212 Views
films has been avoided in order not to deteriorate the surface quality. The composition of layers has been determined by RBS with an accuracy of 1%. Phase identification in the layers with a 0-20 diffractometer revealed the CrSi 2 samples to be single phase, in some layers of f3-FeSi 2 and Ir 3Si5 the X-ray pattern contained traces of other phases. In the case of MnSi 73 the dominant phase was Mn15Si 26 with traces of MnSi. The contamination level of oxygen and carbon in the films, as determined by AES, was between 0.5 and 1 at%. The optical properties between 4 and 24 eV were investigated by spectroscopic ellipsometry which measures both the real and imaginary parts of the dielectric function E and E2, respectively. In our experiments we used synchrotron radiation from the BESSY electron storage ring, between 4 and 9 eV at an angle incidence of 67.50 with a MgF 2 Rochon rotating analyzer and between 13 and 24 eV at an angle of incidence of 450 with a rotating triple-reflection gold analyzer [11]. As the films are polycrystalline, the measured dielectric function is an average over the three principal components of the dielectric tensor. RESULTS AND DISCUSSION The studied semiconducting silicides crystallize in a variety of structures: CrSi 2 hexagonal, Mn, 5 Si 26 tetragonal, fl-FeSi 2 orthorhombic, and Ir 3Si5 monoclinic. Despite this structural diversity the Si-rich silicides share common features. The transition metal lattice is expanded by insertion of an element with s-p valence shells. Bonding in these silicides is governed by the Si-metal interaction. The shortest Si-metal distances are (Cr-Si = 2.47 A, Mn-Si = 2.27 A, Fe-Si = 2.34 A, Ir-Si = 2.335 A).
The degeneracy of metal d states with Si p states favours the formation of strongly covalent bonds leading to bonding and antibonding hybrids. As not all metal d states have the correct symmetry to form hybrids, non-bonding states remain with energies close to those in the element but of reduced band width as the transition metals are pulled farther apart. Therefore, near the Fermi energy the highest valence and the lowest conduction band states have mainly d character, which changes to predominant Si 3s, 3p character for the far-lying valence and conduction bands. Because of this symmetry character interband transition matrix elements are strongly energy dependent. Due to the dominant d character of states on both sides of the gap the across-gap oscillator strengths are very small. Momentum matrix elements become appreciable for states lying farther apart, i. e. for higher energy transitions. In Fig. 1 we show (by points) the results of our absorption measurements at the edge of CrSi 2 which exhibit two absorption steps. The spectral features at the edge were previously assigned to indirect and direct transitions [2,7]. However, recent theoretical calculations of the optical constants of CrSi 2 [12-14] require a revision of this interpretation. The theoretical near-gap absorption data of [12] are shown in Fig. 1 by a solid line. Apart from a shift to higher e
Data Loading...
