Interpretation of structure and magnetism in transition-metal pnictides M 2X and (M 1− xM′ x) 2X

Abstract

The transition-metal pnictides M 2X and (M 1− x M′ x ) 2X containing first-row transition elements and X = P, As or Sb tend to crystallize in three related structures that permit metal-metal bonding via partially filled 3 d-shell cores. It is argued that in the phosphides, and probably in most arsenides and antimonides, of the first-row transition elements, the X-atom p bands are filled and the cation 4 s bands are empty, so that the number of 3 d electrons per metal atom are known unambiguously. Furthermore, some of the phosphides are magnetic and some are not, so that the width of the 3 d-electron bands can be varied by As substitution or by hydrostatic pressure to provide critical information about changes in magnetic order, magneticordering temperatures, and the magnitudes of the atomic moments as a function of bandwidth and band occupation in the critical region where the transition from spontaneous magnetism to Pauli paramagnetism occurs. General conceptual phase diagrams are developed from physical arguments about the influence of electron-electron correlations on quasidegenerate, narrow d bands. This discussion, which leads to an explanation of the Slater-Pauling curve of magnetization vs electron/atom ratio in the transition metals, is then applied to an interpretation of available magnetic data for the transition-metal pnictides M 2X and (M 1− x M′ x ) 2X. Prediction of individual atomic moments requires that allowance be made for distinguishable cation sites and for the transfer of 3 d-electron charge from lighter to heavier elements, but the crystal-field effects appear to be manifest only in the signs of the interatomic exchange interactions. Significantly, in alloys an equal and integral number of majority-spin electrons tend to be stabilized at each atomic constituent that is magnetic.