The dependence of the energy ylso and the halfwidth Slso of 180" L-domain walls in antiferromagnets on the magnetic field H is calculated for low fields. At H + Ho (Ho is the critical field of the spin-flop transition) the walls expand and their energy yls0 decreases. I n the spin-flop region 90" domain structure appears. yso and 6 , are determined by the value of fourth-order anisotropy constant K,. The calculation of the metastable state ( K , < 0) region shows that its boundaries may be approximated by the astroid at lK21 < Kl only. The wall displacement in the spin-flop region ( K , < 0) and the domain structure at K , > 0 are discussed. The problem of the new magnetic phase nucleation is considered. System of 180" Domains in Low FieldsIt is possible that the formation of the antiferromagnetic domain structure is not related with some dominating mechanism, as in the case of ferromagnets (stray fields). It is simply to assume t h a t i t appears by chance at T, [l]. The ground state of antiferromagnets is degenerated, and domains with different orientation of the antiferromagnetic vector 1 are physically equivalent. Two neighbouring domains that differ by the orientation of 1 demand the formation of a domain wall with the shape determined by the competition between shortrange and long-range interactions. It is evident that in the uniaxial antiferromagnets the system of domains is divided by 180" walls [Z]. We consider the domain structure of uniaxial antiferrornagnets in magnetic fields as a system of L-domains divided by 180" walls. The width and the energy of walls should depend on the temperature T and the magnetic field H . The domain structure is destorted when H i s non-parallel to the crystal axes (along which the vector 1 is oriented a t H = 0 ) .

No abstract

The domain structure in the region of the metamagnet transition is considered and the magnetic susceptibility, magnetostriction, and BE-effect are calculated. It is ahown that the appearing domain walls may be of two types: discrete walls having the smaller energy, where an activation energy is necessary to displace them and continuous walls whose energy is greater, but which can be displaced elastically. Two nucleation mechanisms are considered.

Magnetic phase diagrams, magnetization curves, and domain structures of a uniaxial two-sublattice metamagnet are considered taking into account the magnetostatic energy. The problem of nucleation for a metamagnetic phase transition of the first-order type is investigated. The results are compared with available experimental data. PaCCMOTpeHbI MarHHTHbIe $a30BbIe HIlarPaMMbI, KPHBbIe HaMarHHUXBaHHH H HOMeHHbIe CTPYKTYPbI OAHOOCHOrO HByXIIO~pelIIeTOrHOrO MeTaMarHeTHKa IlpH YreTe MWHHTO-CTaTElseCKOH 3HeprMH. HCCJIeJXYeTCZI npo6ne~a 3apO~bI~eO6pa30BaHHH IIpll MeTaMarHHT-HOM @BOBOM IlepeXOHe nepBOr0 pojla. Pe3YJIbTaTbI CpaBHHBaWTCH C Ei MeIoLUHMHCH 3KCIIePHMeHTaJIbHbIMH AaHHbIMH. FeCI, -2 H,O [El, in Gd,Al[13], in rare-earth metals [14], in Gd,Co and (Gd0,2Tb0,8)3Co [15], in Ca,Mn,Ge,O,, [16]. The felaxation of magnetization near OPT is connected apparently with the metamagnetic domain structure [ 171. The metamagnetic OPT are discovered in non-collinear CoNb,O, [18], in four-sublattice PbFeC1, -2 D,O, and CsFeC1, 2 D,O, [19], in rare-earth metal alloys [ZO]. The influence of the concentration z on the metamagnetism of the solution Fe~,-,,Co,Cl, is measured in [21].The authors of a recent experimental study [16] try to interpret the results using the model of a n isotropic metamagnet [22]. However, the agreement with the theory is not complete as the authors of [16] note. I n the present paper we attempt to show that it is possible to explain the available experimental material in the frame of the 33'

No abstract