Temperature memory and resistive glassy behaviors of a perovskite manganite
ABSTRACT This paper reports the observations of long-time relaxation, aging, and temperature memory behaviors of resistance and magnetization in the ferromagnetic state of a polycrystalline La0.7Ca0.3Mn0.925Ti0.075O3 compound. The observed glassy dynamics of the electrical transport appears to be magnetically originated and has a very close association with the magnetic glassiness of the sample. Phase separation and strong correlation between magnetic interactions and electronic conduction play the essential roles in producing such a resistive glassiness. We explain the observed effects in terms of a coexistence of two competing thermomagnetic processes, domain growth and magnetic freezing, and propose that hole-doped perovskite manganites can be considered as "resistive glasses". Comment: Submitted to PRB
- [Show abstract] [Hide abstract]
ABSTRACT: Critical behavior of Ti-doped La0.7Ca0.3MnO3 ceramics are studied using magnetization methods. The results show that the paramagnetic–ferromagnetic transition is first order for the undoped sample, then for low-doping samples the mean-field model is suitable, and finally a 3D Heisenberg model is satisfied for higher doping samples. These findings demonstrate that the critical behavior of the magnetic transition for manganites is sensitive to Mn-site doping.Fuel and Energy Abstracts 01/2010; 322(2):242-246.
- [Show abstract] [Hide abstract]
ABSTRACT: The magnetic and electrical transport properties of La0.9Mn0.9M0.1O3 (M=Mn, Zn and Ti) were investigated. The temperature and magnetic field dependence of electrical resistivity (ρ) and dc magnetization were studied. All the compounds are found in rhombohedral structure. The excess oxygen in all three compounds was detected through iodometric titration. A modification in resistivity is observed when M=Mn is replaced by M=Zn and Ti. The high temperature resistivity above TC follow variable range hopping model for both Zn and Ti compounds. For Zn doping, the observation of large field-cool effect and decrease in resistivity at room temperature and is assumed to be due to the implant of Mn4+ in Mn3+ matrix, which favor Mn3+/Mn4+ double exchange. The ferromagnetic behavior below TC for the compound with M=Ti is correlated to the excess oxygen in it, which implants Mn4+ and thus incorporates ferromagnetic interactions. The substitutions lead to a reduction of Tc and magnetization.Physica B Condensed Matter 07/2012; 407(13):2442–2446. · 1.28 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: A magnetic memory effect was observed by means of zero-field-cooled magnetization measurements in a temperature range below TC. The magnitude of magnetic memory effect, determined by the difference between the magnetization curve with a pause of temperature during the cooling process and its reference curve without the pause, showed that it is independent of the number of pauses. The memory effect is supposed to be related to the isolated states with an "infinite" degeneration, which is associated with the presence of magnetic frustration in a system with a strong magnetic phase competition. Besides, a resistance memory effect was concomitantly observed at the same pause temperatures.Journal of Physics Conference Series 01/2009; 187(1).
arXiv:0802.2729v1 [cond-mat.mtrl-sci] 19 Feb 2008
Temperature memory and resistive glassy behaviors of a perovskite manganite
D. N. H. Nam,∗N. V. Khien, N. V. Dai, L. V. Hong, and N. X. Phuc
Institute of Materials Science, VAST, 18 Hoang-Quoc-Viet, Hanoi, Vietnam
(Received February 19, 2008)
This paper reports the observations of long-time relaxation, aging, and temperature mem-
ory behaviors of resistance and magnetization in the ferromagnetic state of a polycrystalline
La0.7Ca0.3Mn0.925Ti0.075O3 compound. The observed glassy dynamics of the electrical transport
appears to be magnetically originated and has a very close association with the magnetic glassi-
ness of the sample. Phase separation and strong correlation between magnetic interactions and
electronic conduction play the essential roles in producing such a resistive glassiness. We explain
the observed effects in terms of a coexistence of two competing thermomagnetic processes, domain
growth and magnetic freezing, and propose that hole-doped perovskite manganites can be considered
as ”resistive glasses”.
PACS numbers: 72.20.Pa, 72.20.My, 75.47.Lx, 75.50.Lk
Glassy systems are well known for their nonequilib-
rium slow dynamics such as the long-time relaxation,
aging, and memory behaviors. Although the dynamics
of electrons is generally considered very rapid, glassy
transport behaviors have been observed in various sys-
tems including electron glasses in Anderson-insulator film
structures,1,2a 2D electron system in Si3and meso-
porous Si,4ultrathin and granular metal films,5,6rare
earth hydrides YH3−δ,7and a number of perovskite man-
ganite and cobaltite polycrystalline bulks,8,9,10,11single
godic electron dynamics is obviously not expected in con-
ventional electronic systems and the underlying physics
may not be unique considering the fact that it can be
observed in a variety of materials. As for mixed-valence
manganites, R1−xAxMnO3(R: trivalent are earth, A: di-
valent alkaline elements), the slow resistance relaxation
and aging phenomena were noticed quite early by Hel-
molt et al.15after the discoveryof the colossal magnetore-
sistance (CMR) effect.19,20The observed glassy transport
has been commonly attributed to the slow evolution of
the phase conversions among competing phases coexist-
ing in the material as a result of phase separation. Re-
markably, using high-resolution transmission electron mi-
croscopy technique, Tao et al.10observed a dynamic com-
petition between charge-ordered and charge-disordered
phases in La0.23Ca0.77MnO3and related it to the change
in the time dependence of the resistivity.
the authors also suggested that the resistivity relaxation
could be a property of the charge-ordered (orthorhom-
bic) phase itself and not related to the phase coexistence.
Kimura et al.13compared the transport and magnetic
behaviors of Cr-doped Nd1/2Ca1/2MnO3with the dielec-
tric phenomena in ”relaxor ferroelectrics” and proposed
that the system can be viewed as a ”relaxor ferromag-
net”. In addition, Wu et al.12observed glassy transport
in La1−xSrxCoO3cobaltites and specifically ascribed the
phenomenon to the dynamics of the sample’s spin-glass
Such a noner-
We report here the observations of long-time relax-
ation, aging, and memory behaviors of resistivity, which
apparently very much resemble the characteristics of spin
glasses, in the ferromagnetic (FM) state of a polycrys-
talline La0.7Ca0.3Mn0.925Ti0.075O3compound. The resis-
tive glassy dynamics exists in zero magnetic field but is
found to be magnetically originated. Strikingly, tempera-
ture memory behaviors of both resistivity and magnetiza-
tion are revealed in the ferromagnetic state. We explain
the results in terms of two competitive thermomagnetic
processes: domain growth and magnetic freezing. We
propose that glassy transport is a common behavior of
mixed-valence perovskite compounds where phase sepa-
ration occurs spontaneously and the interplay between
magnetism and electrical property is strong; the materi-
als therefore can be considered as ”resistive glasses”.
The La0.7Ca0.3Mn0.925Ti0.075O3 compound was pre-
pared by a conventional solid-state reaction method.
Pure (≥99.99%) raw powders with appropriate amounts
of La2O3, CaCO3, MnO2, and TiO2 were thoroughly
ground, mixed, pelletized and then calcined at several
processing steps with increasing temperatures from 900
oC to 1200oC and intermediate grindings and pelleti-
zations. The product was then sintered at 1300oC for
48 h in ambient atmosphere. The final sample was ob-
tained after a very slow cooling process from the sintering
to room temperature with an annealing step at 700oC
for 5 hours. Room-temperature x-ray diffraction mea-
surements showed that the sample is single phase of a
perovskite orthorhombic (space group Pnma) structure.
Field-cooled (FC) and zero-field-cooled (ZFC) magneti-
zation and four-probe resistivity measurements were car-
ried out in a Quantum Design PPMS-6000. The sam-
ple used for resistivity measurements was a 1.9×1.2×7.5
mm3rectangular bar that was firmly glued to the PPMS
puck to have a good thermal contact but electrically sep-
arated from the puck by a sheet of cigarette paper. To
0 50 100150 200 250300
M (10-1 µB/f.u.)
ρ (Ω cm)
2h at Ta=55K
reference heating (ρref)
50 5560657075 80
ρ (105 Ω cm)
∆ρ (103 Ω cm)
FIG. 1: (color online).
H = 100 Oe) and ρ(T) (right axis, H = 0 and 6 T) of
La0.7Ca0.3Mn0.925Ti0.075O3. (b) ρ(T) measured by different
protocols: ρref(T) was measured on heating following a con-
tinuous cooling from 300 K to 50 K, ρc(T) was recorded dur-
ing cooling the sample to 50 K with a pause at 55 K for 2 h
and ρh(T) was then measured on reheating. The inset of (b)
shows the reduced resistivity ∆ρ(T) = ρref(T) − ρh(T) that
exhibits a maximum at ∼58 K.
(a) MZFC(T), MFC(T) (left axis,
reduce Joule heating and current induced effects, a very
small dc current I = 100 nA (corresponding to a current
density of 4.4 µA/cm2) driven in AC mode was used.
A low heating and cooling rate of 1 K/min was always
chosen for all of the measurements.
III.RESULTS AND DISCUSSION
A.Magnetic and transport characterizations
According to the double-exchange (DE) mechanism for
mixed-valence manganites,21the transfer of an eg elec-
tron between two adjacent Mn ions occurs only when
their localized t2gmagnetic moments are aligned in par-
allel, implying that metallic conducting behavior should
come along with a FM state. Nevertheless, manganites
are strongly separated systems, so that the ideal DE cor-
relation between magnetism and electrical transport may
not be always obtained. While La0.7Ca0.3MnO3is a typ-
ical DE ferromagnet that has a metallic conducting be-
havior in the FM phase below Tc ≈ 250 K, since Ti4+
ions do not take part in magnetic couplings, substitu-
tion of Mn by Ti destabilizes the network of Mn ions,
leading to a deterioration of both DE ferromagnetism
and conductivity.22,23Temperature dependent magneti-
zations, MZFC(T) and MFC(T), and resistivity, ρ(T),
in Fig. 1(a) show that the La0.7Ca0.3Mn0.925Ti0.075O3
compound is an insulator ferromagnet with Tc ≈ 108
K. The insulating behavior in the FM state would sug-
gest that the sample is segregated into metallic conduct-
ing FM regions separated by non-FM insulating bound-
aries. With decreasing temperature, the resistance in-
creases continuously and goes beyond the limit of our
measurement system below 50 K. Under the influence of
a magnetic field H = 6 T, the transport behavior remains
that of an insulator but the resistance is strongly sup-
pressed leading to a negative magnetoresistance, −MR =
(RH=0−RH=6T)/RH=6T, which increases with lowering
temperature and reaches up to ∼9000% at 50 K. The
negative magnetoresistance is probably due to an expan-
sion of the DE FM metallic conducting regions that is
more favored at lower temperatures.
B.Temperature memory and resistive glassy
The ρ(T) data in Fig. 1(a) were measured on reheating
the sample from 50 K following a continuous cooling from
300 K. The ρ(T) curve in zero field is replotted in Fig.
1(b) as a reference heating curve, denoted as ρref(T). In
other measurements, instead of cooling down the sample
continuously directly to 50 K, the cooling was paused at a
constant temperature Ta< Tcfor a duration time ta. In-
terestingly, we observed a clear resistance relaxation that
had no tendency to stop even after several hours during
the pause, showing a signature of an aging process. The
cooling was then resumed to cool the sample to 50 K.
A typical ρ(T) curve recorded during this cooling proce-
dure with Ta= 55 K and ta= 2 h are plotted as ρc(T) in
Fig. 1b, which shows a step at Tacaused by the aging.
Right after reaching 50 K, the temperature was swept
back and ρh(T) data were collected. Intuitively, curves
ρref(T) and ρh(T) are not the same, indicating a cooling
history dependence. By subtracting ρh(T) from ρref(T)
we obtained a temperature dependence of the reduced
resistivity, ∆ρ(T), that shows a maximum at ∼58 K [the
inset of Fig. 1(b)], which is very close to the temperature
where the pause was made. This feature seems to imply
that the pause at Taduring cooling was imprinted into
the system and can be retrieved on reheating, signaling
the presence of a temperature memory effect that was
previously observed in magnetic glassy systems.24,25,26
Henceforth, we refer the measurement of a heating curve
following a paused cooling as ”memory measurement” (or
”memory curve”). In the case where one pause is made,
the memory characterized by a maximum of ∆ρ(T) is
then called ”single temperature memory” (SME) and ∆ρ
is denoted as ∆ρTa
Observations of magnetic glassy behaviors similar to
5055 6065707580 85 90
∆ρSME=ρref-ρH (103 Ω cm)
ρ (Ω cm)
ρ (Ω cm)
ρ (105 Ω cm)
∆ρSME (102 Ω cm)
FIG. 2: (color online). Single temperature memory curves,
55, 60, 65, 70, 75, and 80 K. The arrows mark the ∆ρTa
maximum with the corresponding Ta. The insets display ρ(t)
curves measured at Ta = 60, 100, and 160 K. (b) ∆ρTa
for Ta = 80, 90, and 110 K. The arrows mark the ∆ρTa
curve with the corresponding Ta.
SME(T), with different Ta’s. (a) ∆ρTa
SME(T) curves for Ta =
those of canonical spin glasses in DE perovskite cobaltites
and manganites have been reported in a number of
publications.27,28,29,30However, although signatures of
glassy transport such as slow relaxation and aging were
sometimes observed, memory behaviors of resistance
have not been explored so far. To confirm the memory
behavior in Fig. 1(b), we carried out a series of simi-
lar experiments with ta= 2 h but various temperatures
Ta’s; the results are presented in Fig. 2. The maximum
SMEnear Tais reproducibly obtained but strongly
suppressed with increasing Ta and finally vanished for
Ta> Tc; i.e., it is well defined for Ta= 55, 60, 65, 70,
and 75 K, but appears less and less defined with further
increasing Taand becomes unresolvable for Ta= 110 K.
Irrespective of Ta, ∆ρTa
SMEis strongly suppressed with
temperature and basically becomes zero in the param-
agnetic state. Although a magnetic field is not required
for the observation of this memory effect, these results
unambiguously indicate that it is indeed magnetically
originated. Another noticeable feature is the tendency
SMEat T < Tato change its sign to negative when
Taapproaches close to Tc. Resistivity relaxation curves,
ρ(t), measured during the pauses at Ta= 60 K (≪ Tc),
60 6570758085 9095100
∆ρSME, ∆ρDME (103 Ω cm)
FIG. 3: (color online). Double temperature memory as a com-
bination of the corresponding single temperature memories,
(T) = ∆ρTa1
curve shows a maximum near Ta2 = 60 K and a shoulder
(marked by an arrow) near Ta1 = 70 K. ta1 = 2 h. Symbols:
for ta2 = 2 h: ◦ = ∆ρ70K
▽ = ∆ρ70K
SME(T) + ∆ρTa2
SME(T). The ∆ρTa1,Ta2
SME, ? = ∆ρ60K
SME; for ta2 = 0.5 h: ? = ∆ρ70K,60K
SME, ? = ∆ρ70K,60K
100 K (≈ Tc), and 160 K (≫ Tc) plotted in the insets of
Fig. 2(a) clearly show a very strong downward relaxation
at 60 K, a weaker upward relaxation at 100 K, and no
relaxation at all at 160 K. Qualitatively, the changes in
both amplitude and sign of ρ(t) are consistent with the
variation of the ∆ρTa
SMEcurves. Although the underlying
physics is probably different, this behavior is reminiscent
of the change in relaxation direction of the field-cooled
magnetization in the vicinity of the glass phase transition
in spin glasses and random magnets.25,30
The results in Fig.2 clearly demonstrate that the
system memorizes the temperature where an event was
made during cooling and recalls it on reheating. To figure
out whether the system can memorize separately more
than one event in a single cooling (thus yielding a ”multi-
ple temperature memory”), we made two pauses with the
same ta= 2 h sequently at Ta1= 70 K and Ta2= 60 K in
the same cooling process. The corresponding ∆ρ curve
(Fig. 3) shows a large maximum near 60 K and a small
shoulder around 70 K, indicating a ”double temperature
memory” (DME) behavior. Moreover, this DME is in
fact a combination of the two SME’s at the correspond-
ing Ta’s; i.e., the summation of the two SME curves with
Ta= 60 K and 70 K, ∆ρ60K
tively, matches very well the DME curve ∆ρ70K,60K
as illustrated in Fig. 3. Just for another example, we
reduced ta of the pause at 60 K to 0.5 h in both the
SME and DME experiments and again observed that
the superposition of two SME’s gives the corresponding
DME. Unambiguously, the system can memorize multi-
SME(T) and ∆ρ70K
ple events separately created on a single cooling and store
them without interference – a behavior that has been well
observed in magnetic glasses.24,25,28,29
The use of temperature as the only parameter to vary
the dynamics of the system allows probing the system by
its original dynamics. The observed nonequilibrium be-
haviors of resistance bear striking resemblances with the
dynamic characteristics of magnetic glasses. The resis-
tance keeps relaxing towards equilibrium in such a long
period of time that could go beyond the laboratory time
scale even in a constantly stabilized condition. After be-
ing ”aged” at the pauses during cooling, the system re-
sponds differently on reheating, demonstrating a clear
age-dependent behavior. Furthermore, the temperature
memory effect, an inherent characteristics of magnetic
glasses, is also observed. Since the resistive glassy be-
haviors are observed only in the FM state, they must
have a magnetic origin. The insulating behavior in FM
state suggests that the sample is not a pure DE ferro-
magnet, but probably a separated system consisting of
two competing phases: DE FM metallic-conducting do-
mains separated by a non-FM insulating matrix. Low
temperatures or high magnetic fields both favor the DE
interaction and expand the FM domains at the expense of
the non-FM insulating matrix. During the pause below
Tc, the FM domains would evolve with time by polarizing
the spins surrounding to expand the conducting regions.
On the other hand, thermal activation helps the domain
moments relax towards directions which are energetically
favored by their randomly distributed local anisotropy,
therefore increasing the spin scattering of electrons trav-
eling between domains. The domain growth seems to be
dominant at low temperatures, resulting in a downward
resistivity relaxation (Fig. 2). In contrast, magnetic mo-
ments relax faster at higher temperatures, overcoming
the domain growth process at temperatures close to Tc
and therefore producing an upward resistivity relaxation.
We adopt the N´ eel’s relaxation law31to qualita-
tively describe the freezing of FM domains, τ
τ0exp(Ea/kBT), where Eais the anisotropy energy that
is proportional to the domain size, kBis the Boltzmann’s
constant, and τ and τ0 are the relaxation and micro-
scopic relaxation times, respectively. Lowering temper-
ature leads to an exponential increase of τ, causing the
domains to become deeper frozen. One of the key ingredi-
ents of the temperature memory effect is the freezing of
magnetic moments in random directions that preserves
the ”aged” magnetic configuration when cooling is re-
sumed after the pause. The frozen moments would still
relax but with very much smaller rates at lower temper-
atures. Upon reheating, the melting of the aged con-
figuration returns the system to the state without aging
above Ta, exhibiting the observed memory effect. Since
after the pause the magnetic moments relaxed into direc-
tions with higher energy barriers [kBTaln(ta/τ0), accord-
ing to the N´ eel’s law] than kBTa, higher temperatures
are then required to remove them from the freezing di-
rections. This explains why the maximum of ∆ρ(T) is
∆ρSME (103 Ω cm)
Ta = 65 K
ta = 2 h
FIG. 4: The influence of magnetic field on the nonequilibrium
resistance dynamics. A magnetic field Ha is turned on dur-
ing the pause at Ta = 65 K on cooling. The application of
Ha causes a ”crosstalk” between the memory established at
Ta and the domain configurations developed during further
cooling. The memory maximum in the ∆ρ65K
unresolvable at Ha = 3 kOe.
SME(T) curve is
always at a higher temperature than Ta. In a real system,
there always exist wide distributions of τ0and Ea. Mag-
netic moments of domains with Ea≤ kBTarotate freely
under the thermal excitation, those with Ea≫ kBTaare
effectively frozen, and only those with Ea ? kBTa con-
tribute to the relaxation at Ta. This explains why we
could observe a multiple temperature memory. Ideally,
if Ta2ln(ta2/τ0) < Ta1, there could be no interference be-
tween the two aged configurations. ∆ρ(T) approaches
zero only at sufficiently high temperature far above Ta
where there is effectively no difference in domain sizes
between the memory and the reference configurations.
In order to examine the influence of magnetic field on
the memory effect, a magnetic field Ha was turned on
during the pause.Right after the aging time ta, Ha
was removed and the cooling was immediately resumed.
The SME ∆ρ(T) curves measured with different Ha’s
are presented in Fig. 4. The field is expected to favor
the domain growth and therefore to increase the ∆ρ(T)
maximum near Ta. However, the freezing of magnetic
moments are also strongly affected. Instead of relaxing
towards random directions, the magnetic moments are
polarized and relax preferably towards the external field
direction, establishing a configuration that is advanta-
geous for the unfrozen domains to grow such that the
domain growth during further cooling could effectively
overwhelm the aged configuration previously established
at Ta. We can see clearly in Fig. 4 that, while the max-
imum of the ∆ρ(T) curves is indeed raised with increas-
ing Ha, it is significantly broadened and spreads towards
lower temperatures. Apparently, higher Ha applied at
Ta causes a stronger development of magnetic domains
upon further cooling, which may interfere with the aged
configuration pre-established at Ta. With Ha= 3 kOe,
the maximum of ∆ρ(T) is no longer defined, suggesting
that the development of magnetic domains with decreas-
0 204060 80100 120140
MZFC (10-2 µB/f.u.)
Ta = 65 K
ta = 2 h
H = 10 Oe
H = 30 Oe
50 6070 80
H = 150 Oe
MZFC (10-2 µB/f.u.)
H = 70 Oe
FIG. 5: (color online). The temperature memory behavior
observed on MZFC(T). For the memory curves (symbols), the
sample was zero-field-cooled from 300 K to 5 K with a pause
at Ta = 65 K for ta = 2 h. For the reference curves (lines), the
sample was cooled directly to 5 K. After the temperature had
reached 5 K, a probing field H was immediately applied and
MZFC(T) was recorded on reheating. Main figure: H = 10
Oe, insets (a), (b), and (c): H = 30, 70, and 150 Oe, respec-
tively. Sufficiently high probing fields can align the moments
of the aged domains toward its direction, turning the mem-
ory ”dip” lying below into a ”cusp” lying above the reference
ing temperature below Tabecomes entirely dominant.
C.Temperature memory behavior of
Since the memory behavior of resistance is the result
of thermomagnetic dynamic processes, it is natural to
expect that similar effects can be observed in conven-
tional magnetic measurements, even though the sample
is in a ferromagnetic state. The results in Fig. 5 ver-
ify the temperature memory behavior of magnetization.
With similar cooling procedures as those used for the
transport measurements, depending on the probing field
strength, the SME MZFC(T) curve exhibits a dip or a
cusp around Ta(Fig. 5 insets). For very small fields, the
MZFC at a given temperature is contributed from FM
domains having energy barriers lower than kBT that can
rotate toward the field direction. The freezing of mag-
netic moments that relaxed into higher energy barriers
at Taduring the pause on cooling is thus reflected by a
dip in the memory curve lying below the reference curve.
Melting these moments with increasing T above Tadrives
the memory magnetization toward the reference one. The
effect of larger domain sizes achieved by the pause is not
exhibited because their moments are frozen in random
directions. However, for high probing fields, the rotation
of frozen moments towards the field direction becomes
significant; the larger FM domain sizes in the memory
configuration (than in the corresponding reference one)
give extra contributions to the magnetization. As a re-
sult, with increasing probing field, the memory dip is
gradually shallowed and develops into a cusp lying above
the reference curve . This influence of magnetic field on
the memory behavior is thus in agreement with the pic-
ture of coexisting domain growth and magnetic freezing.
Nonequilibrium dynamic behaviors such as long-time
relaxation, aging, and temperature memory of both re-
sistivity and magnetization have been revealed in the
insulator ferromagnet La0.7Ca0.3Mn0.925Ti0.075O3. The
glassy transport has a very close association with the
magnetic glassiness and appears to be magnetically orig-
inated. These results can be qualitatively explained in
terms of two competing thermodynamic processes: do-
main growth and magnetic freezing. The results of our
dynamical study evidence the presence of phase separa-
tion and strong correlation between magnetism and elec-
trical transport in the CMR perovskite manganite. We
propose that resistive glassiness is a common character
of mixed-valence perovskite manganites as a direct re-
sult of both electronic and magnetic phase separation.
This nonequilibrium dynamics may be found profound
in systems where there exists a strong competition be-
tween conducting and insulating phases, most likely near
the metal-insulator transition with doping.
temperature memory behavior of resistance is associated
with the freezing of magnetic domains, it is expected
to be observable in the manganite insulators where FM
metallic domains are embedded in the non-FM insulating
This work has been performed using facilities of the
State Key Labs (IMS, VAST).
∗Electronic address: email@example.com
1A. Vaknin, Z. Ovadyahu, and M. Pollak, Phys. Rev. B 65,
2V. Orlyanchik and Z. Ovadyahu, Phys. Rev. B 75, 174205
3J. Jaroszy´ nski and D. Popovi´ c, Phys. Rev. Lett. 99, 216401
4S. Borini, L. Boarino, and G. Amato, Phys. Rev. B 75,
5G. Martinez-Arizala, C. Christiansen, D. E. Grupp, N.
Markovi´ c, A. M. Mack, and A. M. Goldman, Phys. Rev. B
57, R670 (1998).
6E. Bielejec and W. Wu, Phys. Rev. Lett. 87, 256601 (2001).
7M. Lee, P. Oikonomou, P. Segalova, T. F. Rosenbaum, A.
F. Th. Hoekstra, and P. B. Littlewood, J. Phys.: Condens.
Matter 17 L439 (2005).
8M. Roy, J. F. Mitchell, and P. Schiffer, J. Appl. Phys. 87,
9P. Levy, F. Parisi, L. Granja, E. Indelicato, and G. Polla,
Phys. Rev. Lett. 89, 137001 (2002).
10J. Tao, D. Niebieskikwiat, M. B. Salamon, and J. M. Zuo,
Phys. Rev. Lett. 94, 147206 (2005).
11K. De, S. Majumdar, and S. Giri, J. Appl. Phys. 101,
12J. Wu, H. Zheng, J. F. Mitchell, and C. Leighton, Phys.
Rev. B 73, 020404(R) (2006).
13T. Kimura, Y. Tomioka, R. Kumai, Y. Okimoto, and Y.
Tokura, Phys. Rev. Lett. 83, 3940 (1999).
14A. Anane, J. -P. Renard, L. Reversat, C. Dupas, P. Veillet,
M. Viret, L. Pinsard, and A. Revcolevschi, Phys. Rev. B
59, 77 (1999).
15R. v. Helmolt, J. Wecker, T. Lorenz, and K. Samwer, Appl.
Phys. Lett. 67, 2093 (1995).
16D. Casa, B. Keimer, M. v. Zimmermann, J. P. Hill, H. U.
Habermeier, and F. S. Razavi, Phys. Rev. B 64, 100404(R)
17X. J. Chen, H. -U. Habermeier, and C. C. Almasan, Phys.
Rev. B 68, 132407 (2003).
18A. Bhattacharya, M. Eblen-Zayas, N. E. Staley, A. L. Ko-
brinskii, and A. M. Goldman, Phys. Rev. B 72, 132406
19R. M. Kusters, J. Singleton, D. A. Keen, R. Mcgreevy, and
W. Hayes, Physica B 155, 362 (1989).
20S. Jin, T. H. Tiefel, M. McCormac, R. A. Fastnacht, R.
Ramesh, and L. H. Chen, Science 264, 413 (1994).
21C. Zener, Phys. Rev. 82, 403 (1951).
22X. Liu, X. Xu, and Y. Zhang, Phys. Rev. B 62, 15112
23D. N. H. Nam, L. V. Bau, N. V. Khiem, N. V. Dai, L. V.
Hong, N. X. Phuc, R. S. Newrock, and P. Nordblad, Phys.
Rev. B 73, 184430 (2006).
24K. Jonason, E. Vincent, J. Hammann, J. P. Bouchaud, and
P. Nordblad, Phys. Rev. Lett. 81, 3243 (1998).
25T. Jonsson, K. Jonason, and P. Nordblad, Phys. Rev. B
59, 9402 (1999).
26R. Mathieu, P. J¨ onsson, D. N. H. Nam, and P. Nordblad,
Phys. Rev. B 63, 092401 (2001).
27D. N. H. Nam, K. Jonason, P. Nordblad, N. V. Khiem, and
N. X. Phuc, Phys. Rev. B 59, 4189 (1999).
28D. N. H. Nam, R. Mathieu, P. Nordblad, N. V. Khiem,
and N. X. Phuc, Phys. Rev. B 62, 8989 (2000).
29R. Mathieu, P. Nordblad, D. N. H. Nam, N. X. Phuc, and
N. V. Khiem, Phys. Rev. B 63, 174405 (2001).
30D. N. H. Nam, R. Mathieu, P. Nordblad, N. V. Khiem,
and N. X. Phuc, J. Magn. Magn. Mater. 226-230, 1340
31N´ eel, Ann. Geophys. (C.N.R.S.) 5, 99 (1949).