Dynamics of the Antiferromagnetic Spin-Flop Transition

Abstract

The critical magnetic field producing spin flop in many antiferromagnets is too small to soften a three-dimensional magnon, i.e., to remove the energy barrier between the equal-energy phases. The barrier can be bypassed, however, via the softening of surface magnons in a smaller field, forming two-dimensional surface-spin-flop states, which broaden with increasing field and which catastrophic ally spread inward across the three-dimensional material as the critical field is approached.It has been suggested that many first-order phase transitions are associated with normal modes which partially but not fully "soften," i.e., which almost go to zero frequency. 1 A nonzero energy barrier between two states of equal free energies provides restoring forces confining the normal-mode oscillations to one side of the barrier and to finite frequencies. In those first-order transitions without superheating or supercooling (i.e., without hysteresis) there arises the problem of how the system manages to hurdle the energy barrier. We exhibit here a possible mechanism for one such system.As first shown theoretically by N6el 2 and experimentally by Poulis and Hardeman, 3 at a critical magnetic field H 3 (see Table I) applied along the preferred axis of an antiferromagnet (AF) there occurs a sudden, nearly 90° rotation of the sublattice vectors into the so-called "spin-flop" (SF) phase. The AF-SF phase transition is first order, there being a finite jump in the magnetic moment, and it can readily be made to take place near 0 K. One's first thought is that it occurs at that field iJ 4 at which the AF resonance frequency goes to zero, i.e., where the restoring torques vanish and a mode softens. Early work either assumed H 3 =H 4 or attempted to derive it. 4 Anderson and Callen 5 identified the three possible critical fields H 2 , H 3 , and H 4 described in Table I. They argued that the AF-SF transition should occur at H 4 in increasing H but at H 2 in decreasing H; and that unless H 2 = H 3 =H 4 , hysteresis should be observed.Experimental evidence of the absence of hystereresis has been accumulating. 6 Furthermore, the AF -SF transition in at least three materials 7 takes place at a field measurably less than H 4 and theoretically evaluated as H 3 , i.e., in the presence of a finite energy barrier.Mills and Saslow 8 have shown that a field H 1 <H 3 fully softens a surface magnon; Mills 9 has demonstrated that above this same H t there can occur surface-spin-flop (SSF) regions, in which a few layers of spins near a surface have turned nearly 90°. According to Mills, the number of layers involved decreases with increasing H. We show that Mills's analysis omitted terms of the same order of magnitude as those included, and that in fact the number of layers increases with increasing H.Our picture of the dynamics of the AF-SF transition is as follows: At a field H l surface magnons soften and SSF regions develop. Similar localized regions might also develop near impurity clusters, dislocation lines, and other imperfections. The SSF region...

Surface spin-flop transition in an antiferromagnet

“…¢ÑÎÇÇ ÕÑÅÑ, ÍÂÍ ÃÞÎÑ ÒÑ-ÍÂÊÂÐÑ Ä ÓÂÃÑÕÇ [2], àÕÑ ÒÓÇAEÒÑÎÑÉÇÐËÇ âÄÎâÇÕÔâ ÑÛË-ÃÑÚÐÞÏ, ÒÑÔÍÑÎßÍÖ ÄÃÎËÊË ÒÑÄÇÓØÐÑÔÕË ÄÑÊÐËÍÂÇÕ ÐÇ 90 -ÐÂâ,  180 -ÐÂâ ÑÃÏÇÐÐÂâ ÔÒËÓÂÎß. £ ÕÇÚÇÐËÇ ÏÐÑÅËØ ÎÇÕ ÒÑÔÎÇ ÒÑâÄÎÇÐËâ ÓÂÃÑÕ [4,5] ÄÑÒÓÑÔ Ñ ÒÑÄÇÓØÐÑÔÕÐÑÏ ÔÒËÐ-×ÎÑÒ-ÒÇÓÇØÑAEÇ ÑÔÕÂÄÂÎÔâ ÄÐÇ ÊÑÐÞ ÄÐËÏÂÐËâ ËÔÔÎÇAEÑÄÂÕÇÎÇÌ.¯ÑÄÞÌ ËÐÕÇÓÇÔ Í ÐÇÏÖ ÄÑÊÐËÍ ÒÑÔÎÇ ÑÕÍÓÞÕËâ âÄÎÇÐËâ ÅËÅÂÐÕÔÍÑÅÑ ÏÂÅÐÇ-ÕÑÔÑÒÓÑÕËÄÎÇÐËâ Ä ÏÐÑÅÑÔÎÑÌÐÞØ ÏÂÅÐËÕÐÞØ ÔÕÓÖÍÕÖ-ÓÂØ (ÔÏ., ÐÂÒÓËÏÇÓ, ÓÂÃÑÕÞ [6 ë 8] …”

Observation of an atomic exchange bias effect in DyCo<inf>4</inf> film

“…The lower atomic coordination of the atoms at the surface results in a band narrowing, leading to changes of the moments at the surface in respect to the bulk. In MnF 2 19,20 , it requires a lower magnetic field value to force the phase transition from an AF to a spin-flop state at the surface. Ferromagnetism on the surface of the bulk paramagnetic intermetallic compound YCo 2 was predicted using full-potential electronic structure calculations 21 and observed by spin-resolved photoelectron spectroscopy 22 .…”

Spin Waves in Antiferromagnetically Coupled Multilayers

“…In particular, these effects have been studied in Fe/Cr systems of different type [1][2][3]. From the theoretical point of view, several investigations have been devoted to the study of the static and dynamical magnetic properties of antiferromagnetically coupled multilayers with a finite even numbers N of ferromagnetic layers [4][5][6][7].Within the mean field approach proposed by Mills [8], and lately adopted by several authors [9,[4][5][6][7], the energy of N magnetic layers in an applied magnetic field Ho is given by (in magnetic-field units):where 1 is the layer index and HE and HA (both positive) are the effective interlayer exchange coupling and the uniaxial anisotropy constant, respectively. The angle αl between the magnetization of the l-th layer and the external field Ho is positive for clockwise rotation with 1 odd and negative with 1 even.…”

“…Within the mean field approach proposed by Mills [8], and lately adopted by several authors [9,[4][5][6][7], the energy of N magnetic layers in an applied magnetic field Ho is given by (in magnetic-field units):…”

“…In all the cases we have investigated, the ground state of the system has been determined by minimizing the energy given by Eq. (1) following the procedure outlined by Keifer and Chow [9].…”