PHYSICAL REVIEW E 86, 041107 (2012)

Deep spin-glass hysteresis-area collapse and scaling in the three-dimensional ± J Ising model
Ozan S. Sarıyer,1 Alkan Kabakçıoğlu,1 and A. Nihat Berker2,3
1
Department of Physics, Koç University, Sarıyer 34450, Istanbul, Turkey
2
Faculty of Engineering and Natural Sciences, Sabancı University, Orhanlı, Tuzla 34956, Istanbul, Turkey
3
Department of Physics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA
(Received 31 May 2012; published 5 October 2012)
We investigate the dissipative loss in the ±J Ising spin glass in three dimensions through the scaling of the
hysteresis area, for a maximum magnetic field that is equal to the saturation field. We perform a systematic analysis
for the whole range of the bond randomness as a function of the sweep rate by means of frustration-preserving
hard-spin mean-field theory. Data collapse within the entirety of the spin-glass phase driven adiabatically (i.e.,
infinitely slow field variation) is found, revealing a power-law scaling of the hysteresis area as a function of the
antiferromagnetic bond fraction and the temperature. Two dynamic regimes separated by a threshold frequency
ωc characterize the dependence on the sweep rate of the oscillating field. For ω < ωc , the hysteresis area is
equal to its value in the adiabatic limit ω = 0, while for ω > ωc it increases with the frequency through another
randomness-dependent power law.

DOI: 10.1103/PhysRevE.86.041107 PACS number(s): 05.70.Ln, 75.60.Ej, 64.60.Ht, 75.10.Nr

Hysteresis in magnetic materials has been a subject of H in the second term in Eq. (1) is the uniform external mag-
interest for quite some time due to its applications in magnetic netic field. With a proper choice of units, the temperature for
memory devices and as a testing ground for theories of the system may be defined as T ≡ 1/J . A random distribution
nonequilibrium phenomena [1–4]. The hysteresis area which of the ferromagnetic and antiferromagnetic bonds gives rise
measures the magnetic energy loss in the material is connected to frustration and yields a spin-glass phase for a range of p
with the Barkhausen noise [5,6] due to irreversible avalanche values. Ising spin-glass models are widely used as a tool for
dynamics [7–12]. The existing literature on hysteresis in ran- understanding the properties of experimental spin glasses such
dom magnets focuses mostly on random-field models [12–15] as Pr0.6 Ca0.4 Mn0.96 Ga0.04 O3 [11], Fe0.5 Mn0.5 TiO3 [25–27],
while numerical studies on random-bond models are mostly at LiHo0.167 Y0.833 F4 [28], and Cu3−x AlMnx [29]. Without loss
zero temperature [16–22]. To our knowledge, there has been of generality we set p  0.5 since the partition function is
no finite-temperature study of the hysteresis loss, especially invariant under the transformation p,{siA },{sjB } → (1 − p),
in the spin-glass phase where large avalanches are expected to {siA },{−sjB }, where A and B signify the two sublattices.
be severely prohibited. We here investigate the adiabatic and For small values of p and H = 0, the orientational (up-
dynamic hysteresis in the the ±J random-bond Ising spin glass down) symmetry is spontaneously broken below a critical
[23] on a finite, three-dimensional simple cubic lattice with temperature Tc (p) and long-range ferromagnetic order sets
periodic boundary conditions. We show that the hysteresis area in. This phase is well understood within the Landau picture
obeys a scaling relation in the whole spin-glass phase, in accord where the free energy landscape is described by two minima
with earlier theoretical studies which observed scale invariance at magnetizations ±m(T ,p). Beyond a critical fraction pc of
over the whole range about the critical disorder for various the antiferromagnetic bonds, reducing temperature drives the
disorder-driven systems [15–17]. Moreover, this scaling data system into a glassy phase. The low-temperature phase now
collapse is also observed for experimental systems over wide retains its orientational symmetry and a new, randomness-
ranges of the temperature and the magnetic field: Gingras dominated phase which has a broken replica symmetry appears
et al. observed a universal data collapse over four decades [30,31]. In this phase, the free energy landscape is rough, with
in a geometrically frustrated antiferromegnet Y2 Mo2 O7 [24], many local minima at significantly nonoverlapping configu-
while Gunnarsson et al. observed such a data collapse for the rations. Meanwhile, the dynamics slows down to the extent
short-range Ising spin glass Fe0.5 Mn0.5 TiO3 [25]. that the relaxation time diverges [32]. At high temperatures
The ±J Ising spin-glass model is defined by the dimen- T > Tc (p), both ordered phases give way to a paramagnetic
sionless Hamiltonian state where the entropic contribution to the free energy is
dominant. While the critical temperature strongly depends on
  p along the ferromagnet-to-paramagnet phase boundary, only
−βH = Jij si sj + H si , (1)
a weak dependence of Tc on p is observed for the spin-glass
ij  i
phase [32,33]. In this study, we investigate the hysteretic
behavior of a spin glass under the uniform magnetic field H
where β ≡ kB1T is the inverse temperature. The first sum that is swept at a constant rate ω. A past computational study
in Eq. (1) is over the pairs of nearest-neighbor sites (i,j ), similar to ours [34] considered a time-dependent quenched-
where Jij is the quenched-random local interaction between random magnetic field that was conjugate to the spin-glass
the classical Ising spins si = ±1. The probability distribution order parameter.
function for Jij is given by We use hard-spin mean-field theory (HSMFT), a self-
consistent field theoretical approach [34–50] that preserves
P (Jij ) = p δ(Jij + J ) + (1 − p)δ(Jij − J ) . (2) the effects due to the frustration (crucial for the spin-glass

1539-3755/2012/86(4)/041107(6) 041107-1 ©2012 American Physical Society

SARIYER, KABAKÇIOĞLU, AND BERKER PHYSICAL REVIEW E 86, 041107 (2012)

phase) generated by the randomly scattered antiferromagnetic 1 .0

bonds. HSMFT is defined by the refined set of self-consistent
equations 0 .5
   
  

m
0 .0
mi = P (mj ,sj ) tanh Jij sj + H (3)
− 0 .5
{sj } j j

for the local magnetization mi at each site i, whose nearest − 1 .0 T = 4.00 T = 2.00
7500
p = 0.05 p = 0.05
neighbors are labeled by j . The single-site probability distri- p = 0.15 p = 0.22
bution is 1000 p = 0.35 p = 0.35 5000
1 + mj sj

tR
P (mj ,sj ) = . (4)
2 500 2500

The local magnetization mi at site i satisfies −1  mi  1.

The hard-spin mean-field theory Eq. (3) has been discussed in 0 0
− 0 .1 0 .0 − 0 .1 0 .0 0 .1
detail by the authors of Refs. [34–50]. H
HSMFT has been successfully applied to spin glasses
[34,43]. In this paper we make use of the method to investigate FIG. 1. (Color online) Hysteresis curves (upper) and relaxation
the scaling of the hysteresis area under a uniform, time- times (lower) at high (T = 4.00, left) and low (T = 2.00, right) tem-
dependent magnetic field. To this end, we consider a 20 × peratures. Data are for p values either deep in the ferromagnetic (p =
20 × 20 cubic lattice with periodic boundary conditions. We 0.05), spin-glass (T = 2.00, p = 0.35), or paramagnetic (T = 4.00,
have checked in this study and in a previous study [34] that our p = 0.35) phases, or close to the phase boundaries for the
hard-spin mean-field theory results are independent of size for ferromagnetic-paramagnetic (T = 4.00, p = 0.15) or ferromagnetic-
an L × L × L system for L  15. A particular realization at a spin-glass (T = 2.00, p = 0.22) transitions. For each case, an overlay
given (T ,p) is generated by the assignment of the quenched- of 20 distinct runs with different random-bond arrangements is shown.
random coupling constants Jij according to the probability
The para-ferro and para-spin-glass phase boundaries are easily
distribution of Eq. (2) and, initially, a random and unbiased
determined by locating the temperature at which A0 vanishes
choice of spins si = ±1. To determine the hysteresis curves,
(i.e., falls below m ). A set of p scans for different temperatures
the system is first saturated by a sufficiently large external
and a set of temperature scans for various p values are given
field Hs , the minimum value of H for which Eq. (3) yields
in Fig. 3. The low-temperature ferro-spin-glass boundary is
an average magnetization m = (1/L3 ) i mi = 1 within an
located at pc  0.22 and is calculated as the inflection point
accuracy m ≡ 10−6 . Then, the path Hs → −Hs → Hs is
for the maximum slope of the hysteresis curve as a function of
traversed with steps H = Hs /100 or smaller. For each
antiferromagnetic bond probability [16]. The phase boundaries
incremental change of the field, the system is allowed to relax
are consistent with the well-known phase diagram for the
a number of time steps τ = 1/ω. A time step corresponds to
three-dimensional ±J model [33] and in fair comparison with
successive iterations of Eq. (3) on L3 arbitrarily chosen sites.
An infinitely slow sweep is obtained as the limit τ → tR , where
the HSMF equations converge to a self-consistent solution 5.5 100
within the tolerance interval m . Thus, tR is the relaxation time
5.0
of the system. 10−1
The infinitely slow-sweep hysteresis curves obtained in 4.5 PM
the ferromagnetic and spin-glass phases are shown in Fig. 1. 10−2
The usual jump in the magnetization at a coercive field 4.0
T = 1/J

Hc , observed for small p, is associated with a system-wide 10−3

avalanche in the ferromagnetic phase. For p larger than a 3.5
critical value pc , this picture is replaced by a slanted hysteresis 10−4
curve and a smaller hysteresis area, typical of spin-glass ma- 3.0
FM SG
terials [3,11,29]. This converse hysteretic behavior, associated 10−5
2.5
with the Barkhausen noise [5,6], is a consequence of the
power-law distribution of avalanches which is well established 2.0 10−6
[6,7,10–12,14–18,20,21,29,51] for several frustrated systems 0.0 0.1 0.2 0.3 0.4 0.5
p
with quenched disorder. The hysteresis area disappears in the
paramagnetic phase. FIG. 2. (Color online) Logarithmic contour plot of the infinitely
In Fig. 2, we present the infinitely slow-sweep hysteresis slow-sweep hysteresis area A0 as a function of antiferromagnetic
area globally, for all temperatures and antiferromagnetic bond probability p and temperature T = 1/J . The thick vertical line
bond probabilities, on a logarithmic color-contour plot. The denotes the phase boundary between the ferromagnetic and the spin-
hysteresis area A0 vanishes in the region shown in dark glass phases as described in the text, while the other thick line bounds
blue, which corresponds to the paramagnetic phase, while the paramagnetic phase where the infinitely slow-sweep hysteresis
it is nonzero in the ferromagnetic and spin-glass phases, area is less than the precision used in the consistent-field calculations
respectively, on the left and right of the lower half of Fig. 2. (i.e., A0 < 10−6 ).

041107-2
DEEP SPIN-GLASS HYSTERESIS-AREA COLLAPSE AND . . . PHYSICAL REVIEW E 86, 041107 (2012)

100 0.10 100 0.50

0.45
0.15
10−1 10−1
0.40
0.20
−2 0.35
10−2 10
0.25 0.30
10−3 0.25

A0
−3 0.30
A0

10
0.35 0.20
10−4 10−4 0.15
0.40
0.10
10−5 10−5
0.45 0.05
10−6 0.50 10−6 0.00
0.0 0.1 0.2 0.3 0.4 0.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
p T = 1/J

FIG. 3. (Color online) Infinitely slow-sweep hysteresis area A0 , as a function of antiferromagnetic bond probability p for temperatures
(indicated in the color legend) 1/T = J = 0.10, 0.11, . . . ,0.50 (left) and as a function of temperature T = 1/J for antiferromagnetic bond
probabilities (indicated in the color legend) p = 0.00, 0.01, . . . ,0.50 (right). Each curve is a tenth degree polynomial fit to the averages over
20 realizations.

the experimental temperature-concentration phase diagrams increases with increasing sweep rate ω. This can be understood
of the various Eux Sr1−x Sy Se1−y , solid (o-H2 )1−x (p-H2 )x , and by observing that the slow response of the magnetization
AuFe systems reviewed in Ref. [32]. to a time-varying field inflates the hysteresis curve along
We here focus on the scaling form of the hysteresis area in the field direction. The typical behavior observed in various
the spin-glass phase and show that a unique scaling-function experimental and theoretical magnets (typically pure magnets
governs the whole range of p and J within the spin-glass phase. or random-field systems) [52–56] is
To this end, we first express the hysteresis area in the form
A0 = A0 (p̃,J˜), where p̃ ≡ p−p c
and J˜ ≡ J −J c
are the reduced A(ω,p,J ) = A0 + g(p,J ) ωb , (8)
pc Jc
displacements from phase boundaries. We then postulate the
where b is the sweep-rate exponent. We investigate whether
multivariate scaling form
the random-bond Ising spin glass obeys a similar scaling
A0 (p̃,J˜) = λc A0 (λa p̃,λb J˜) , (5) relation.
A typical scan of the hysteresis area as a function of ω
which by letting λ = p̃−1/a reduces to displays two dynamic regimes, separated by a critical sweep
A0 (p̃,J˜) = p̃−c/a A0 (1,p̃−b/a J˜) , (6)
3
Defining ν ≡ c/a, μ ≡ −b/a, and f (x) ≡ A0 (1,x), we obtain × 10 −
15 0.50
p̃ν A0 (p̃,J˜) = f (p̃μ J˜) . (7) 10 − 2 0.47
The sought collapse is obtained by the choice of scaling 12 10 − 3
exponents μ = 1 and ν = 2. The data shown in Fig. 3 collapse 10 − 4 0.44
onto a single curve shown in Fig. 4, where the left-hand 10 − 5
1 . 72 0.41
side (LHS) of Eq. (7) is plotted against the argument on the 9 10 − 6
right-hand side (RHS) for 28 evenly spaced values of p above 10 − 7 0.38
p̃ ν A 0

pc . The origin corresponds to the phase boundary between 10 − 3 10 − 2 10 − 1 10 0

0.35
the spin-glass and paramagnetic phases. The log-log plot 6
of the same collapse shown in the inset of Fig. 4 suggests 0.32
that the scaling function has the form f (x) ∝ x 1.72 , yielding
a hysteresis area A0 ∝ p̃α J˜β with α  −0.28 and β  1.72. 3 0.29
Interestingly, unlike the case of the usual critical phenomena,
0.26
the scale-invariance applies to the entire spin-glass phase and
not just to the vicinity of the critical phase boundary. 0 0.23
Having analyzed the limit with infinitely slow-sweep rate, − 0 .9 − 0 .6 − 0 .3 0. 0 0. 3 0 .6
we next consider the dynamic hysteretic response as a function
p̃ μ J˜
of the magnetic field frequency. One can simulate the finite
oscillation frequency by iterating Eq. (3) for a predetermined FIG. 4. (Color online) Scaling of the hysteresis area in the
number of steps t, instead of waiting until a steady state spin-glass phase as a function of reduced antiferromagnetic bond
is reached. The sweep rate ω = 1/t is proportional to the concentration p̃ and the reduced bond strength J˜, for various p values
frequency of the applied field up to a material-dependent spin as shown in the color legend. The scaling function f (x) given by the
relaxation time. The hysteresis area A(ω,p,J ) deviates from RHS of Eq. (7) on which all data points collapse is consistent with a
the value at infinitely slow sweep A0 = A(ω = 0,p,J ) and same power law within the entire spin-glass phase.

041107-3
SARIYER, KABAKÇIOĞLU, AND BERKER PHYSICAL REVIEW E 86, 041107 (2012)

0.0 0.1 0.2 0.3 0.4 0.5

T = 2.0
1.2 T = 1.0
T = 0.5
100
1.0

b
0.8
10−2

0.6

10−4 0.4
T = 2.0
0.0 0.1 0.2 0.3 0.4 0.5
p
10−6
FIG. 6. (Color online) Sweep-rate exponent b versus antifer-
0
romagnetic bond fraction p for temperatures T = 2.0, 1.0, and
10 0.5. The dashed curve depicts the general trend of the sweep-rate
exponent, while the dotted vertical line marks the phase transition
from ferromagnetic to spin-glass phase.

10−2 by an incremental increase in the field decay within a period

A − A0

1/ω or smaller. For faster sweeps (ω > ωc ), the increase in

the area follows the power law in Eq. (8), with a p-dependent
exponent b. In the ferromagnetic phase with weak disorder,
the two dynamic regimes are separated by a sharp increase in
10−4 the hysteresis area. This transition gets significantly smoother
T = 1.0 in the spin-glass phase, especially far from the ferromagnetic-
spin-glass boundary. For larger systems, one expects ωc to
recede and the power-law behavior to dominate.
10−6 Figure 6 shows the sweep-rate exponent b calculated as
a function of the antiferromagnetic bond fraction p, at fixed
100 temperatures T = 1/J = 2.0, 1.0, and 0.5. The hysteresis area
is calculated for the sweep rates ω = 1, 0.5, 0.3̄, 0.2, 10−1 , . . . ,
10−4 at each p value, after averaging over ten realizations.
The exponent values are obtained through fits to the data
in the regime ω > ωc (typically two decades or more), using
10−2 the functional form of Eq. (8). The error bars reflect only the
scatter of the data relative to the fit. In the ferromagnetic phase
p < pc , we note that the calculated sweep-rate exponents
lie in an interval of fairly good agreement with the various
10−4 values obtained previously at p = 0, namely b = 2/3 [52–55]
and b = 0.52 ± 0.04 [53] from mean-field theory, b = 0.61
T = 0.5
[53] from Glauber dynamics simulations, b = 0.495 ± 0.005
[54] and b = 0.45 [56] from Monte Carlo simulations.
In conclusion, we have considered here the ±J Ising
10−6 −4
10 10−3 10−2 10−1 100 101 model under a uniform external field and investigated the
ω scaling behavior of the saturation hysteresis area (i.e., far
from the weak-field limit). We observed that the phase
FIG. 5. (Color online) Hysteresis area difference A − A0 versus diagram can be derived from the hysteresis area alone and
sweep rate ω, for temperatures T = 2.0, 1.0, 0.5 from top to bottom the ferromagnetic-spin-glass phase boundary corresponds to
and for antiferromagnetic bond fractions p = 0.0, 0.1, . . . ,0.5 as the inflection point with regard to bond-randomness strength
shown in the color legend. p. When adiabatically driven, the area displays a data collapse
within the entire spin-glass phase for all temperatures and p.
rate ωc that depends on p, J , and the system size (Fig. 5). For The scaling function itself has a power-law form and the scale
a sufficiently slowly varying field ω < ωc , the area is pinned at invariance extends far from the phase boundary, deep into the
the value A0 . In this regime, the avalanches that are triggered spin-glass phase.

041107-4
DEEP SPIN-GLASS HYSTERESIS-AREA COLLAPSE AND . . . PHYSICAL REVIEW E 86, 041107 (2012)

The dynamical response under a fluctuating external field the increase in the hysteresis area with ω is due to the magnet’s
is also interesting. We find that, beyond a threshold value ωc , delayed response to the changing field, and a signature of
the hysteresis area increases as a function of the field-sweep the spin-glass phase is the slowing down of precisely such
rate ω with a nonuniversal power law. This behavior is not relaxation phenomena.
limited to the vicinity of the phase transition. The associated
exponent is found to be a function the randomness strength p.
Moreover, this function is independent of temperature. In the ACKNOWLEDGMENTS
limit of a pure magnet (p → 0), we observe good agreement
with the existing literature, despite the fact that the earlier Support by the Alexander von Humboldt Foundation, the
theoretical work applied to a weak driving field, while we here Scientific and Technological Research Council of Turkey
consider sweeps across saturation limits. Figure 6 suggests (TÜBİTAK), and the Academy of Sciences of Turkey (TÜBA)
that, relative to the ferromagnetic phase, the spin glass displays is gratefully acknowledged. We acknowledge the hospitality
an amplified sensitivity to the field-sweep rate, again running of the TÜBİTAK-Bosphorus University Feza Gürsey Institute
in apparent contrast to the general wisdom that the hysteretic for Fundemental Sciences, for the computational support from
effects are suppressed within a spin glass. In fact, we note that the Gilgamesh cluster.

[1] G. Bertotti, Hysteresis in Magnetism: for Physicists, Materials [22] H. G. Katzgraber, F. Pázmándi, C. R. Pike, K. Liu, R. T. Scalettar,
Scientists, and Engineers (Academic, Amsterdam, 1998). K. L. Verosub, and G. T. Zimányi, Phys. Rev. Lett. 89, 257202
[2] F. Colaiori, Adv. Phys. 57, 287 (2008). (2002).
[3] M. S. Pierce, C. R. Buechler, L. B. Sorensen, S. D. Kevan, E. A. [23] G. Toulouse, Commun. Phys.2, 115 (1977), reprinted in Spin
Jagla, J. M. Deutsch, T. Mai, O. Narayan, J. E. Davies, K. Liu, Glass Theory and Beyond, edited by M. Mezard, G. Parisi, and
G. T. Zimanyi, H. G. Katzgraber, O. Hellwig, E. E. Fullerton, M. A. Virasoro (World Scientific, Singapore, 1987), pp. 99–103.
P. Fischer, and J. B. Kortright, Phys. Rev. B 75, 144406 (2007). [24] M. J. P. Gingras, C. V. Stager, N. P. Raju, B. D. Gaulin, and
[4] H. G. Katzgraber and G. T. Zimanyi, Phys. Rev. B 74, 020405(R) J. E. Greedan, Phys. Rev. Lett. 78, 947 (1997).
(2006). [25] K. Gunnarsson, P. Svedlindh, P. Nordblad, L. Lundgren,
[5] H. Barkhausen, Physik Zeits 20, 401 (1919). H. Aruga, and A. Ito, Phys. Rev. B 43, 8199 (1991).
[6] G. Durin and S. Zapperi, in The Science of Hysteresis, edited by [26] H. G. Katzgraber, D. Hérisson, M. Östh, P. Nordblad, A. Ito,
G. Bertotti and I. D. Mayergoyz, Vol. 2 (Academic, Amsterdam, and H. A. Katori, Phys. Rev. B 76, 092408 (2007).
2005), p. 181. [27] A. Ito, H. Aruga, E. Torikai, M. Kikuchi, Y. Syono, and H. Takei,
[7] P. Bak, C. Tang, and K. Wiesenfeld, Phys. Rev. Lett. 59, 381 Phys. Rev. Lett. 57, 483 (1986).
(1987). [28] W. Wu, D. Bitko, T. F. Rosenbaum, and G. Aeppli, Phys. Rev.
[8] P. J. Cote and L. V. Meisel, Phys. Rev. Lett. 67, 1334 (1991). Lett. 71, 1919 (1993).
[9] L. P. Lévy, J. Phys. I (Paris) 3, 533 (1993). [29] E. Vives, E. Obradó, and A. Planes, Physica B 275, 45 (2000).
[10] J. P. Sethna, K. Dahmen, S. Kartha, J. A. Krumhansl, B. W. [30] G. Parisi, Phys. Rev. Lett. 50, 1946 (1983).
Roberts, and J. D. Shore, Phys. Rev. Lett. 70, 3347 (1993). [31] M. Mézard, G. Parisi, N. Sourlas, G. Toulouse, and M. Virasoro,
[11] V. Hardy, S. Majumdar, M. R. Lees, D. McK. Paul, C. Yaicle, J. Phys. (Paris) 45, 843 (1984); Phys. Rev. Lett. 52, 1156 (1984).
and M. Hervieu, Phys. Rev. B 70, 104423 (2004). [32] K. Binder and A. P. Young, Rev. Mod. Phys. 58, 801 (1986).
[12] O. Perković, K. Dahmen, and J. P. Sethna, Phys. Rev. Lett. 75, [33] Y. Ozeki and H. Nishimori, J. Phys. Soc. Jpn. 56, 1568 (1987).
4528 (1995); O. Perković, K. A. Dahmen, and J. P. Sethna, Phys. [34] B. Yücesoy and A. N. Berker, Phys. Rev. B 76, 014417
Rev. B 59, 6106 (1999). (2007).
[13] S. Sabhapandit, P. Shukla, and D. Dhar, J. Stat. Phys. 98, 103 [35] R. R. Netz and A. N. Berker, Phys. Rev. Lett. 66, 377 (1991).
(2000). [36] R. R. Netz and A. N. Berker, J. Appl. Phys. 70, 6074 (1991).
[14] K. A. Dahmen, J. P. Sethna, M. C. Kuntz, and O. Perković, [37] J. R. Banavar, M. Cieplak, and A. Maritan, Phys. Rev. Lett. 67,
J. Magn. Magn. Mater. 226, 1287 (2001). 1807 (1991).
[15] J. P. Sethna, K. A. Dahmen, and O. Perković, in The Science [38] R. R. Netz and A. N. Berker, Phys. Rev. Lett. 67, 1808 (1991).
of Hysteresis, edited by G. Bertotti and I. D. Mayergoyz, Vol. 2 [39] R. R. Netz, Phys. Rev. B 46, 1209 (1992).
(Academic, Amsterdam, 2005), p. 107. [40] R. R. Netz, Phys. Rev. B 48, 16113 (1993).
[16] E. Vives and A. Planes, Phys. Rev. B 50, 3839 (1994). [41] A. N. Berker, A. Kabakçıoğlu, R. R. Netz, and M. C. Yalabık,
[17] E. Vives, J. Goicoechea, J. Ortı́n, and A. Planes, Phys. Rev. E Turk. J. Phys. 18, 354 (1994).
52, R5 (1995). [42] A. Kabakçıoğlu, A. N. Berker, and M. C. Yalabık, Phys. Rev. E
[18] E. Vives and A. Planes, J. Phys. IV (Paris) 5, C2-65 (1995). 49, 2680 (1994).
[19] A. K. Hartmann, Phys. Rev. B 59, 3617 (1999). [43] E. A. Ames and S. R. McKay, J. Appl. Phys. 76, 6197 (1994).
[20] E. Vives and A. Planes, J. Magn. Magn. Mater. 221, 164 [44] G. B. Akgüç and M. Cemal Yalabık, Phys. Rev. E 51, 2636
(2000). (1995).
[21] F. Pázmándi, G. Zaránd, and G. T. Zimányi, Physica B 275, 209 [45] J. E. Tesiero and S. R. McKay, J. Appl. Phys. 79, 6146 (1996).
(2000). [46] J. L. Monroe, Phys. Lett. A 230, 111 (1997).

041107-5
SARIYER, KABAKÇIOĞLU, AND BERKER PHYSICAL REVIEW E 86, 041107 (2012)

[47] A. Pelizzola and M. Pretti, Phys. Rev. B 60, 10134 (1999). [52] P. Jung, G. Gray, R. Roy, and P. Mandel, Phys. Rev. Lett. 65,
[48] A. Kabakçıoğlu, Phys. Rev. E 61, 3366 (2000). 1873 (1990).
[49] H. Kaya and A. N. Berker, Phys. Rev. E 62, R1469 (2000); also [53] G. P. Zheng and J. X. Zhang, J. Phys.: Condens. Matter 10, 1863
see M. D. Robinson, D. P. Feldman, and S. R. McKay, Chaos (1998).
21, 037114 (2011). [54] G. P. Zheng and J. X. Zhang, Phys. Rev. E 58, R1187 (1998).
[50] T. Çağlar and A. N. Berker, Phys. Rev. E 84, 051129 [55] G. P. Zheng and M. Li, Phys. Rev. B 66, 054406 (2002).
(2011). [56] M. Acharyya and B. K. Chakrabarti, Phys. Rev. B 52, 6550
[51] E. Vives and A. Planes, Phys. Rev. B 63, 134431 (2001). (1995).

041107-6