3. Phase Transition Phenomena of Diluted Antiferromagnets in a Magnetic Field and Random-Field Effects

Abstract

Our present work gives an unified interpretation for the phase transition behaviors in three-dimentional dilute Ising antiferromagnets under a uniform magnetic field. Untii now, theoretical and experimental studies of phase transition in random magnetic fields have been extensively performed. The issue of the lower critical dimention d_⊥ of Ising system in random fields has attracted much theoretical and experimental interest. While Imry and Ma originally suggested that d_⊥=2 by a simple domain-wall argument, and calculations valid for dimensions d>4 suggest d_⊥=3, there has been still a controversy. Fishman and Aharony first pointed out that random.fields can be generated in uniaxial random antiferromagnets by applying an external field. Using a random bond model, they showed that a Hamiltonian of a two-sublattice antifcrromagnet in a uniform field is equivalent to that of a ferromagnet in random fields. Random fields arc generated indirectly by the random molecular fields and their mngnetudc is linearly proportional to the local magnetization. Thereafter, modification of random fields is carried out by using a random site model, and they are also generated directly by the applied field. It has been demonstrated by various experimental results that in a two-dimensional system, random magnetic fields destroy the long-range-order at low temperature and the phase transition. On the other hand, in three-dimensional systems, an interpretation for various experimental results on neutron scattering, specific-heat and magnetization measurements has been still controversial; some suggest that the phase transition is a first order, some assert that d_⊥=2, and others suggest that d_⊥≧3. Accordingly, we have performed comprehensive studies of magnetic susceptibility and specific-heat of a three-dimensional dilute antiferromagnet Mn_<0.82>Zn_<0.18>F_2 in a uniform field. We get a new concept from the present experiment that two kinds of magnetic fields are generated in a dilute antiferromagnet in a uniform field. One is, of course, a random field and the other is a staggered field. A staggered field destroys the phase transition and a random field does not. A divergent magnetic susceptibility at extremely-weak fieid arid a rounding of the peak at higher fields directly suggests that a staggered magnetic field is induced by a uniform field. Our observation of the field dependence of the specific-heat is as follows; in smaller fields specific-heat reveals a symmetric divergence with peak height larger than in zero field and upon further increasing a fieid a peak rounds. The phenomenon in small field is attributed to the effect of a random field. But the behaviors that the peaks fall and round with increasing fields are attributed to that of a staggered field. From these considerations, we can say that in the real magnetic systems two compcating fields in dilute antiferromagnets in a uniform field have made an interpretation for the experimental results to be difficult and that random magnetic field dose not destroy the phase transition in three-dimentiorial Ising systems.