Designing Crystal Structures from Atoms up
- PDF / 140,756 Bytes
- 6 Pages / 612 x 792 pts (letter) Page_size
- 1 Downloads / 248 Views
DD8.6.1
Designing Crystal Structures from Atoms up Shahab Derakhshan, Enkhtsetseg Dashjav, and Holger Kleinke* Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1 E-mail: [email protected] ABSTRACT A novel structure map was developed for metal-rich pnictides and chalcogenides M2Q, with M being an element of groups 3 - 5, and Q of groups 15 - 16. To date, this family exhibits eleven different structure types, which are separated in well-defined domains in this structure map. The map comprises a combination of atomic factors as the abscissa, namely principal quantum numbers, valence-electrons, and radii, and a structural factor as the ordinate, which is the averaged coordination number of the Q atoms. We demonstrate in this contribution how this new map can be and has been used to predict crystal structures of new materials from atoms up. First attempts to expand this concept towards different stoichiometries are described. INTRODUCTION Structure maps can provide a guide for the systematic synthesis directed towards new compounds of a known structure type. Probably the best-known ones are the Mooser-Pearson [1], Zunger [2], and Villars [3] maps. All three use exclusively physical coordinates, i.e. electronegativity differences, averaged principal quantum numbers, pseudopotential radii, and valence-electrons. A different approach was taken by Pettifor, who developed a different structure map using phenomenological relative order numbers. These are based on a string running through the periodic table, the path of which was optimized empirically to get the best separations for AB compounds [4]. Unfortunately, all of these maps fail to separate the eleven structure types of the more than 35 examples of the metal-rich pnictides and chalcogenides M2Q of the metal groups 3 - 5. This number includes the ternaries (M,M')2Q and M2(Q,Q') [5-7]. This class continues to grow, as the number of recent publications from different researchers show; even new binaries were continuously found during the last five years, namely Hf2Te [8], Sc2Te [9], β-Ti2Se [10], and Zr2Te [11]. The eleven structure types, though exhibiting quite various structural motifs, share two common features. First, the Q atoms are always surrounded solely by M atoms that are located on the vertices of fragments of deltahedral tetrakaidecahedra; and second, the M atoms form extended substructures (one- to three-dimensional) with numerous M−M bonds. The latter usually are reminiscent of common element structures, such as the hexagonal closed packing (hcp, realized in, e.g. Hf2S [12]), face-centered cubic (fcc, e.g. La2Sb [13]), and, most often, body-centered cubic (bcc, e.g. Ta2P [14]). An exception is Ta2S, whose Ta atom substructure comprises chains of icosahedra [15], while the isovalent Ta2Se exhibits bcc fragments [16]. CONSTRUCTION OF THE M2Q STRUCTURE MAP How can one understand the formation of the different structure types, and even predict, which structures would be adopted by new compounds of this family? Based on the obs
Data Loading...
