COHERENT POTENTIAL APPROXIMATION FOR RANDOMLY DILUTED UNIAXIAL ANTI-FERROMAGNETS - APPLICATION TO MN1-XZNXF2

A recently proposed CPA theory (Tahir-Kheli et al), for treating randomly diluted antiferromagnets with uniaxial anisotropy is revised. It is shown that such a procedure, which ignores dynamical effects occasioned by fluctuations in the local anisotropy field, leads to unphysical results beyond certain critical concentrations at which the theory breaks down. A study of the weakness, the relevance and validity of this theory is presented. The theory is expected to be relevant only when applied to systems with relatively small anisotropy. In fact, owing to its simplicity, within its range of validity, we apply a such CPA to analyse the behaviour of the zero-field antiferromagnetic resonance in the mixed system of manganese and zinc fluorides as function of zinc concentration. Our results are compared with those from several recent experiments and a very good agreement is found. Moreover, the general features and complexity of an alternative CPA formalism which takes into account dynamical effects of fluctuations in both exchange and anisotropy fields, are also briefly discussed.