Real and Hypothetical Intermediate-Valence Ag(II)/Ag(III) and Ag(II)/Ag(I) Fluoride Systems as Potential Superconductors.

With the aim of gauging their potential as conducting or superconducting materials, we examine the crystal structures and magnetic properties of the roughly one hundred binary, ternary, and quaternary Ag(II) and Ag(III) fluorides in the solid state reported up to date. The Ag(II) cation appears in these species usually in a distorted octahedral environment, either in an [AgF](+) infinite chain or as [AgF(2)] sheets. Sometimes one finds discrete square-planar [AgF(4)](2-) ions. The Ag(III) cation occurs usually in the form of isolated square-planar [AgF(4)](-) ions. Systems containing Ag(III) (d(8)) centers are typically diamagnetic. On the other hand, the rich spectrum of Ag(II) (d(9)) environments in binary and ternary fluorides leads to most diverse magnetic properties, ranging from paramagnetism, through temperature-independent paramagnetism (characteristic for half-filled band and metallic behavior) and antiferromagnetism, to weak ferromagnetism. Ag(II) and Ag(III) have the same d-electron count as Cu(II) (d(9)) and Cu(III) (d(8)), respectively. F(-) and O(2-) ions are isoelectronic, closed-shell (s(2)p(6)) species; both are weak-field ligands. Led by these similarities, and by some experimental evidence, we examine analogies between the superconducting cuprates (Cu(II)/Cu(III)-O(2-) and Cu(II)/Cu(I)-O(2-) systems) and the formally mixed-valence Ag(II)/Ag(III)-F(-) and Ag(II)/Ag(I)-F(-) phases. For this purpose we perform electronic-structure computations for a number of structurally characterized binary and ternary Ag(I), Ag(II), and Ag(III) fluorides and compare the results with similar calculations for oxocuprate superconductors. Electronic levels in the vicinity of the Fermi level (x(2)-y(2) or z(2)) have usually strongly mixed Ag(d)/F(p) character and are Ag-F antibonding, thus providing the potential of efficient vibronic coupling (typical for d(9) systems with substantially covalent bonds). According to our computations this is the result not only of a coincidence in orbital energies; surprisingly the Ag-F bonding is substantially covalent in Ag(II) and Ag(III) fluorides. The electron density of state at the Fermi level (DOS(F)) for silver fluoride materials and frequencies of the metal-ligand stretching modes have values close to those for copper oxides. The above features suggest that properly hole- or electron-doped Ag(II) fluorides might be good BCS-type superconductors. We analyze a comproportionation/disproportionation equilibrium in the hole-doped Ag(II) fluorides, and the possible appearance of holes in the F(p) band. It seems that there is a chance of generating an Ag(III)-F(-)/Ag(II)-F(0) "ionic/covalent" curve crossing in the hole-doped Ag(II)-F(-) fluorides, significantly increasing vibronic coupling.