Training-induced inversion of spontaneous exchange bias field on La1.5Ca0.5CoMnO6

0.98 MB·

by  L. Bufaical

Training-induced inversion of spontaneous exchange bias field on La Ca CoMnO
1.5 0.5 6
L. Bufai¸cal,1,∗ R. Finkler,1 L. T. Coutrim,1 P. G. Pagliuso,2 C.
Grossi,3 F. Stavale,3 E. Baggio-Saitovitch,3 and E. M. Bittar3
1Instituto de F´ısica, Universidade Federal de Goia´s, 74001-970, Goiaˆnia, GO, Brazil
2Instituto de F´ısica “Gleb Wataghin”, UNICAMP, 13083-859, Campinas, SP, Brazil
3Centro Brasileiro de Pesquisas F´ısicas, 22290-180, Rio de Janeiro, RJ, Brazil
(Dated: January 16, 2017)
7 In this work we report the synthesis and structural, electronic and magnetic properties of
1 La1.5Ca0.5CoMnO6 double-perovskite. This is a re-entrant spin cluster material which exhibits a
0 non-negligiblenegativeexchangebiaseffectwhenitiscooledinzeromagneticfieldfromanunmag-
2 netizedstatedowntolowtemperature. X-raypowderdiffraction,X-rayphotoelectronspectroscopy
and magnetometry results indicate mixed valence state at Co site, leading to competing magnetic
n
phases and uncompensated spins at the magnetic interfaces. We compare the results for this Ca-
a
J doped material with those reported for the resemblant compound La1.5Sr0.5CoMnO6, and discuss
themuchsmallerspontaneousexchangebiaseffectobservedfortheformerintermsofitsstructural
2
1 andmagneticparticularities. ForLa1.5Ca0.5CoMnO6,whensuccessivemagnetizationloopsarecar-
ried, the spontaneous exchange bias field inverts its sign from negative to positive from thefirst to

thesecondmeasurement. Wediscussthisbehaviorbasedonthedisorderatthemagneticinterfaces,
]
i relatedtothepresenceofaglassyphase. Thiscompoundalsoexhibitsalargeconventionalexchange
c
bias, for which there is no sign inversion of the exchangebias field for consecutive cycles.
s
-
l
r
t I. INTRODUCTION A-site with cations of different charge and/or radii is
m
a straightforward method to obtain interesting proper-
t. The exchange bias (EB) effect is a well known phe- ties as glassy magnetic behavior11, structural/magnetic
a phase separation12, ferroelectricity13 and also conven-
m nomena characterized by a shift of the magnetization
tional and spontaneous exchange bias effects9,10. The
as a function of applied field (H) hysteresis loop along
d- the H axis. It is usually ascribed to the unidirectional La2−xSrxCoMnO6double-perovskite(DP)seriesexhibits
the largest ZEB effect reported so far, for which 25% of
n exchange anisotropy at the interface between antiferro-
o magnetic (AFM) and ferromagnetic (FM)/ferrimagnetic Sr substitution at La site leads to the largest shift on
c (FIM)phasesinheterogeneoussystems1,2,butithasalso theM(H)curve11. Sr2+ substitutionatLa3+ siteonthe
[ parentcompoundLa CoMnO leadstoanti-sitedisorder,
been observed at the interfaces of spin glass (SG) and 2 6
1 AFM/FM/FIM phases3,4. This effect is conventionally mixedvalenceofCoionsandcompetitionbetweendiffer-
v observed after cooling the system through its N´eel tem- ent magnetic phases, resulting in a re-entrant spin-glass
6 perature (T)in the presence of H or if H is applied dur- (RSG) behavior9.
9
ing the materials’s fabrication. Recently, improvements The microscopic mechanisms responsible for the ZEB
4
onoxideheterostructuresgrowthhaverenewedtheinter- effectobservedonDPcompoundsarenotyetunderstood,
3
0 estinsuchcompounds,since itopens the realpossibility and in order to verify the currently proposed ideas it is
. of fabrication of multifunctional devices using strongly important to study novel materials. In this context, two
1
0 correlated electronic materials5. factorsmakeLa1.5Ca0.5CoMnO6 averyinterestingfocus
7 Recently, some compounds have been reported to ex- of investigation. Firstly, there is an advantage in doping
1 hibit a spontaneous EB effect, i.e., there is a shift along La2CoMnO6withCainsteadofSr,whichisthefactCa2+
: the field axis on the magnetizationas a function ofmag- ionic radius is closer to La3+ than Sr2+14. As it is well
v
neticfieldmeasurement[M(H)]evenwhenthesystemis known, DP oxides exhibit a strong correlation between
i
X cooledfromanunmagnetizedstateandwithoutthepres- their structural, electronic and magnetic properties. As
r ence of external magnetic field6–10. Despite the different such, it is always very difficult to predict which of these
a mechanisms claimed to explain this zero field cooled ex- properties would be more strongly affected by doping,
change bias (ZEB) effect on distinct materials, two in- and consequently, would lead the variation of the oth-
gredients emerge as responsible for the phenomena: the ers. The use of Ca, instead of Sr, on La1.5A0.5CoMnO6
presence of a glassy magnetic state re-entrant to a con- leadstopresumablelessstructuralchangesinrelationto
ventionalmagnetismandthereconfigurationofthespins La2CoMnO6, making it easier to notice the influence of
duringtheinitialfieldapplicationattheM(H)measure- electronic changes due ionic substitution. Moreover, the
ment. comparison between the Ca- and Sr-doped compounds
The transition-metal (TM) oxides are potential can- provides insights about the influence of crystal structure
didates to exhibit the ZEB effect due to their intrinsic to the ZEB effect.
structural/magnetic inhomogeneity. From this class of In this work, we report the synthesis and investiga-
materials, the perovskite-type (ABO , A-rare/alkaline- tion of structural, electronic and magnetic properties of
3
earth and B-TM) stands out. Ionic substitution at the the new DP compound La Ca CoMnO . It presents
1.5 0.5 6
2
aclusterspinglass(CG)phaseconcomitantlywithAFM
and FM phases, being thus a RSG material. It also ex-
hibits both the spontaneous and conventional exchange
bias (CEB) effects. Our results are compared to those
reportedfor the analoguecompoundLa Sr CoMnO ,
1.5 0.5 6
and we discuss the differences in terms of structural
and magnetic particularities of each compound. Inter-
estingly, when consecutive M(H) loops are measured on
La Ca CoMnO after cooling the system in the ab-
1.5 0.5 6
sence of applied magnetic field there is a sign inversion
of the ZEB field from negative to positive values from
thefirsttothesecondloop. Wediscussthisresultonthe
basis oftwo contributionsto the EBeffect, the rotatable
and frozen spins at the AFM/FM/CG interfaces.
II. EXPERIMENT DETAILS
FIG. 1: Rietveld refinement fitting of La1.5Ca0.5CoMnO6.
Theverticallinesrepresent theBragg reflections andtheup-
PolycrystallineLa Ca CoMnO wassynthesizedby
1.5 0.5 6 perpanel shows theschematic structure of the DP.
conventionalsolid-statereactionmethod. Stoichiometric
amounts of La O , CaO, Co O and MnO were mixed
2 3 3 4
◦
and heated at 800 C for 12 hours in air atmosphere. respectively)11, it can be noted that Co–O shrinks in
Later the sample was re-grinded before a second step relation to the undoped material while Mn–O remains
◦
of 24 hours at 1200 C. Finally the materialwas grinded, nearly unchanged. The TM valences for La CoMnO
◦ 2 6
pressed into pellets and heated at 1300 C for 24 hours. are believed to be Co2+ and Mn4+, thus the results sug-
Highresolutionx-raypowderdiffraction(XRD)datawas gest that Ca2+ substitution at the La3+ site acts mainly
collectedatroomT usingaBrukerD8Discover diffrac- to change the Co valence. This is corroborated by XPS
tometer, operating with Cu Kα radiation over the angu- measurements,whichindicatemixedvalenceCo2+/Co3+
◦ ◦
lar range 10 ≤ θ ≤ 100 , with a 2θ step size of 0.01 . (see below).
Rietveld refinement was performed with GSAS software Another important result observed is that Co–O–Mn
and its graphical interface program15. X-ray photoelec- bond angle for our Ca-doped sample (158.8◦) is smaller
tronspectroscopy(XPS)experimentswereperformedus- than those observed for the Sr-doped compounds
inganultra-highvacuum(UHV)chamberequippedwith La Sr CoMnO (163.8◦)16, La Sr CoMnO
1.75 0.25 6 1.6 0.4 6
a SPECS analyzer PHOIBOS 150. AC and DC mag- (165.3◦)17 and LaSrCoMnO (180◦)18, which is con-
6
netization measurements were carried out using a com- sistent to the fact that larger A-site ions lead to more
mercial SQUID-VSM magnetometer. DC data measure- symmetric crystal structure for DP compounds19,20.
ments were performed in both zero field cooled (ZFC) Since the magnetic couplings between TM ions are
and field cooled (FC) modes. AC magnetization mea- strongly correlated to the bond angles, this result is
surementswereperformedinoscillatingfieldwithampli- probablyoneofthefactorsresponsibleforthedifferences
tudeHac =5Oe,insixdifferentfrequenciesintherange between the magnetic properties of La1.5Ca0.5CoMnO6
25-10000Hz. and La Sr CoMnO , as will be discussed next.
1.5 0.5 6
III. RESULTS AND DISCUSSION B. X-ray photoelectron spectroscopy
A. Crystal structure All the XPS results here displayed correspond to the
use of monochromatic Al-K x-ray radiation (hν =
α
La Ca CoMnO room T XRD data revealed the 1486.6 eV) using spectrometer pass energy (E ) of 15
1.5 0.5 6 pass
formation of a single phase DP (Fig. 1). Its struc- eV.Thespectrometerwaspreviouslycalibratedusingthe
ture could be successful refined in the monoclinic P2 /n Au 4f (84.0 eV) which results on a full-width-half-
1 7/2
space group, the same observed for the relative com- maximum (FWHM) better than 0.7 eV, for a sputtered
poundLa CoMnO 11. Thisisnotsurprising,sinceCa2+ metallicgoldfoil. PriortomountingthesampleinUHV,
2 6
and La3+ ionic radii are very close in XII coordination it was slightly polished and ultra-sonicated sequentially
(1.34 ˚A and 1.36 ˚A respectively)14. inisopropylalcoholandwater. Inthefollowing,thesam-
The main parameters obtained from the Rietveld re- ple was mounted on the top of a pure gold foil and ref-
finement are displayed in Table I. Comparing the aver- erenced by setting the adventitious carbon C 1s peak to
age Co–O (1.98 ˚A) and Mn–O (1.95 ˚A) bond lengths 284.6 eV.
to those reported for La CoMnO (2.04 ˚A and 1.94 ˚A In the Fig. 2(a) the XPS survey spectra for
2 6
3
TABLE I: Structuralparameters for La1.5Ca0.5CoMnO6 obtained from Rietveld refinement.
a (˚A) b (˚A) c(˚A) β (◦) Co–O (˚A) Mn–O (˚A) Co–O–Mn (◦) χ2 Rwp
5.4843(1) 5.4498(1) 7.7099(2) 89.94(1) 1.981(14) 1.954(14) 158.8(5) 1.7 9.6
ThedetailedanalysisoftheCo2pregionisquiteuseful
to determine the electronic structure of the cation. This
analysis is however far from trivial, and the main peaks
are known to be composed by multiplet splitting as de-
scribed by Gupta and Sen29, and extensively discussed
in several recent studies30. Nevertheless, we have con-
sidered that the intensity of each spin-orbit component
can be fitted using two curves related to the different
CocationsandrespectivesatellitessinceCo2+ andCo3+
are known to display well distinguish binding energies
at about 780.1 eV and 779.6 eV. The spectral analysis
forLa Ca CoMnO employingthetwofittingcompo-
1.5 0.5 6
nentsresultsontheassignmentofCo2+andCo3+species
located at binding energies and FWHM of 780.1 (2.84)
and779.3(1.82)eV,respectively. Thefittedcomponents
can hardly determine the absolute amounts but serves
to indicate the ratio of Co3+/Co2+ cations in the com-
pound, about 1.3.
FIG. 2: (a) XPS survey spectra for La1.5Ca0.5CoMnO6. All
The Mn 2p core level spectrum is shown in Fig. 2(c).
elementsobservedinthesampleandthegoldfoilsupportsare
indicated. (b) and (c) display the high resolution Co 2p and Thesplitting betweenthe 2p3/2 and2p1/2 levelsisabout
Mn2pregionsandthecorrespondingpeakcomponentsfitting 11.5 eV, and are peaking at 641.5 eV and 653.0 eV, re-
usingaShirleybackgroundandtwoLorentzian-Gaussian(30) spectively. Once again, the relative concentration of Mn
peaks components. cationsoxidationstate is difficult to assignbut the spec-
tral features observed are indicative of Mn at either 3+
or4+valencestates31,32. Themainpeaksarefittedwith
La Ca CoMnO is shown and all elements found are two components indicating the relevant peak features.
1.5 0.5 6
indicated, including the gold foil support signal. Due to The presence of a shoulder on the high energy side and
thepolycrystallinenatureofthe materialthecarbonsig- the relatively broad and symmetric peak is allusive of
nal persist on the sample’s surface and any attempt to Mninmanganite33. We believethe Mnspeciesareprob-
gentleArgonsputterthesurfaceresultsonspecieschem- ably in the Mn4+ oxidation state since no evidence of
ical reduction. Anyhow, the XPS still serves as a pow- Mn2+ or Mn3+, such as satellite features, are observed
erful tool to investigate the electronic structure of the and the typical peak asymmetry of Mn4+ is only sup-
cobalt and manganese cations in the compound. Based pressed by the hydroxylation of the surface during sam-
ontherelativelylargeelectronmeanfreepathfortheCo plecleaning34,35. TheXPSresultssupportourXRDand
2p and Mn 2p photoelectrons (larger than 1.5 nm) the magnetometryfindingsonthevalencestateoftheCoand
spectra shown reveals chemical information from buried Mn cations in the compound, which can be interpreted
oxide layers20–25. At the present conditions, although a in terms of Co2+/Co3+ mixed valence state (see below).
detailed analysisofthe cobaltandmanganese2p regions
aredifficult,onecanyetaddresstheCoandMnchemical
state based on the spectral features.
C. AC and DC Magnetization
It has been largely reported in the literature that the
expectedbindingenergiesforCo3+ andCo2+ arelocated
at 779.5 and 780.5 eV, respectively26. Noteworthy, de- Fig. 3(a) shows the ZFC and FC T dependent mag-
netic susceptibility (χ) for La Ca CoMnO , obtained
pendingonthecationsymmetryonemayobserveintense 1.5 0.5 6
satellite features related to Co2+ at octahedral sites, in at H = 100 Oe. For T ≥ 200 K, the data could be well
contrast to Co3+ at octahedral or Co2+ at tetrahedral fitted to the Curie-Weiss law [inset of Fig. 3(a)]. The
sites27. The Co 2p spectra shown in figure 2(b) presents effective magnetic moment obtained from the fit is µeff
= 6.9 µ /f.u., which can be compared to the value esti-
prominentpeaksat779.7and795.0eVrelatedtothe Co B
mated from the usual equation for systems with two or
2p and2p spin-orbitcomponents,respectively. Re-
3/2 1/2 more different magnetic ions10,36
markablehoweveristhelackofsatellitefeatureswhichin
various compounds is a clear signature of Co2+ in octa-
hedral sites, most probably at high-spin configuration28. µ= µ 2+µ 2+µ 2+.... (1)
1 2 3
p
4
Weiss temperature, θCW ∼ 170 K. The positive value

7 (a) mu)24 1/ iinntdeircaactteisonthsaotnththeeFsMystceomu.plBinugtdaommiangantiefisedthveiemwagonfetthiec
e Fit to Curie-Weiss Law
6 ol/18 ZFC curve [Fig. 3(b)] reveals two anomalies above 100
ol) 5 FC -1 (m 12 C =W 6=. 91 67B /Kf.u. Kst,atweh.icThhceoCulod2+b–eOa–ssMocni4a+tedsuwpeitrhextchheaCngoeminixteerdacvtailoenncies
m predicted to be FM39. The 25% of Ca2+ to La3+ substi-
6
u/ 4 tution on La2CoMnO6 induces the emergence of Co3+,
m 0 andtheCo3+–O–Mn4+ couplingisexpectedtobeAFM.
e 3 120 160 200 240 280 Hence the second anomaly observed on Fig. 3(b) could
( T (K) be related to the emergent AFM ordering. It can also
2 be noted a slight change in the slope of the curve at
ZFC H = 100 Oe T ∼ 55 K. It becomes more evident when the derivative
1 of the curve is carried, as indicated on the inset of Fig.
3(b). The presence of competing magnetic phases, to-
0 gether with the disorder that intrinsically accompanies
the DP compounds, are understood as key ingredients
0 50 100 15 0 200 250 300 to induce glassy states. As will be discussed next, this
T (K)K) inflectionpointisrelatedtothe emergenceofaCGstate
ol at low-T.
m
u/ InordertoverifythepresenceofaSG-likephase,which
m
e is believed to play an important rule on the ZEB effect,
ol) 2 dT ( we have measured AC magnetic susceptibility as a func-
m d tionofT forsixfrequencies(f)inthe25-10000Hzrange.
/ Fig. 4showsthereal(χ’)andimaginary(χ”)partsofthe
u
m 60 90 120 150 180 ac susceptibility for some selected frequencies. Fig. 4(a)
T (K)
e showsvariationsatlow-T intheχ’curvesfordifferentf,
(
indicative of glassy behavior. For the χ” curves on Fig.
1 4(b) one can see a peak (T - freezing T) that decreases
f
in magnitude and shifts to higher T as f increases. This
ZFC
is characteristic of SG-like systems40. The evolution of
T with f turns out to be well described by the power
f
(b) law equation of the dynamical scaling theory
50 100 150
−zν
T (K) τ (Tf −Tg)
= (2)
τ (cid:20) T (cid:21)
0 g
FIG.3: (a)ZFCandFCT dependentmagneticsusceptibility.
Theinsetshowstheinverseofmagneticsusceptibilityandits whereτ istherelaxationtime correspondingtothe mea-
fit to the Curie-Weiss law. (b) Magnified view of the ZFC
sured f, τ is the characteristic relaxation time of spin
curveat low-T, and the inset shows its derivative. 0
flip,T isthetransitiontemperature(T asf approaches
g f
zero) and zν denotes the critical exponent40. The solid
line on inset of Fig. 4(b) represents the fit to Eq. 2,
Assuming ∼50%/50% proportion of high spin (HS) yielding T ≃ 46.7 K, τ ≃ 3.1×10−7 s and zν ≃ 8.9,
g 0
Co2+/Co3+ and complete quenching of orbital angular being these values typical of CG systems11,40.
moments for the TM ions located at an octahedral en-
Usually, a formula computing the shift in T between
f
vironment (µ = µ = 3.9 µ , and µ =
Co2+ Mn4+ B Co3+ the two outermostfrequencies is usedto classify the ma-
4.9 µ )37, the estimated magnetic moment is µ = 5.9
B terial as SG, CG or superparamagnet
µ /f.u., which is smaller than the experimental value
B
and may indicate the presence of Mn3+ (µ = 4.9
Mn3+
△T
µB). On the other hand, some orbital contribution for δTf = f . (3)
the moments is usually observed on Co- and Mn-based Tf △logf
perovskites20, which must also be considered. Using the
L+S moments for Mn4+ (4 µB) and HS Co2+ and Co3+ ForCGsystems,δ usuallylies inthe range0.01.δTf .
(5.2 and 5.5 µB, respectively)38 on Eq. 1 yields µ = 0.110,40–42,andforLa1.5Ca0.5CoMnO6wefoundδ ≃0.07.
6.7µB/f.u.,whichisveryclosetotheexperimentallyob- ThisCGphase,togetherwiththe conventionalmagnetic
served value. transitionsobservedathigherT,confirmstheRSGstate
Fromthefitting ofχ−1 itwasalsoobtainedtheCurie- on this material.
5

(a) 3 (a)
100 Hz
ol)0.8 215000 0H Hzz 2 ZFC
m )
u/ 2500 Hz u. 1 2K
em0.6 /f. B H (k Oe)
( 0 -10 -9 9 10
’

M
-1 u.)
0.4 Hac = 5 Oe /f. B
-2 M 0 ->+7
0-->-7
30 60 90 120 150 -3
T (K) -80 -60 -40 -20 0 20 40 60 80
Tf 1.4 T)g (b) H ( kOe)
ol) 2.7 Fit with Eq. 2 1.3 og(T - f (b)
m 1.2 L
/
u
m 1.1 )2.0
2 e -5.5 -5.0 -4.5 -4.0 -3.5 -3.0 .u.
- 0 Log(10-3f ) /f
(1 1.8 B ~37 kOe

’’ Hac = 5 Oe M 1.5
~21 kOe
0.9
20 40 60 80 1.0
T (K)
20 30 40 50
FIG. 4: (a) χ’ac and (b) χ”ac as a function of T for some H (kOe)
selected frequencies. The curves were obtained with Hac=5
Oe oscilating field. The inset shows T as a function of f,
f FIG. 5: (a) ZFC M(H) loop at 2 K. The inset shows a mag-
where the solid line represents the best fit to the data using nified view of the loops measured in the modes 0 → 7T→
Eq. 2. −7T→ 7T and 0 → −7T→ 7T→ −7T. (b) Magnified view
of the intersections between the virgin curve and the second
branch of the loop.
D. Exchange bias
The M(H) curves were measured at various T. In or-
pliedfieldH =70kOe,whichcanbeexplainedbythe
der to prevent the presence of remanent magnetization, max
presence of AFM and CG phases. The large coercivity
it was followed a careful protocol of sending the sample
masks the curves asymmetry for the complete loop, but
tothe paramagneticstateanddemagnetizingthe system
a magnified view on inset of Fig. 5(a) reveals the nega-
in the oscillating mode between one measurement to an-
tive shift of the M(H) curve (i.e., the shift direction is
other. This protocol is particularly important for the
opposite to the positive field initially applied in the vir-
ZFC M(H) measurements, since any trapped current in
gincurve). ToverifythiseffectwehavemeasuredM(H)
the magnet would prevent the correctobservationof the
with the initial H in the opposite (negative) direction.
ZEB effect. Moreover, to get a reliable comparison be-
As can be observed on inset of Fig. 5(a), the curve ex-
tween the results obtained at different T, all measure-
hibitsapositiveshift. Theshiftintheoppositedirection
mentswerecarriedwiththesameexperimentalprotocol,
is an expected behavior of a EB system, hence a clear
i.e., the same scan length, maximum moment range, fre-
evidence that this result is intrinsic of the material.
quency, T-rate on the ZFC procedure etc.
Fig. 5(a) shows the curve measured at 2 K, obtained The ZEB effect is believed to be related to the field
after cooling the sample in ZFC mode. It is a closed induced reconfiguration of the spins at the magnetic in-
loop that does not saturate even for the maximum ap- terfaces that occurs during the initial magnetization of
6

250 ZFC 9 2.5 /f.u.)B2 2K 10
8 0
200 2.0 M
Oe) HZEB 7 Oe) Oe) -2 8 e)
H ( ZEB150 HC 6 H (kC (kEB1.5 -60 -30 H (0kOe)30 60 6 (kOC
- 100 C H
5 -H1.0 HFC = 50 kOe
50
4 0.5 4
HCEB
0 3 0.0 HC
0 5 10 15 20 25 2
5 10 15 20 25
T (K)
T (K)
FIG.6: -HZEBandHC evolutionwithT. Thelinesareguides
for the eye. FIG. 7: -HCEB and HC evolution with T, after cooling the
samplewithHFC =50kOe. Thelinesareguidesfortheeye.
The inset shows theFC M(H) loop at 2 K.
thesystem6–10. ThusintheM(H)curveofaZEBmate-
rialthevirgincurveusuallyliesoutsidethemainloopfor
2740Oe. This signifies a greatenhancement in compari-
a certain H interval, and for some critical H the curves
son to the ZFC measurement, and resembles results ob-
intersect6,7,9. ThiscanbeseenforLa Ca CoMnO on
1.5 0.5 6 served for related compounds as La Sr CoMnO and
1.5 0.5 6
Fig. 5(b), where the virgin curve and the second branch
La Ca CoIrO ,forwhichH inducesthe pinningof
1.5 0.5 6 FC
ofthe hysteresisloopinterceptsatH ∼21kOe. Butdif- the spins already at high T,enhancing the EB effect9,10.
ferentlythanforotherZEBmaterials,atH ∼37kOethe
InEBsystems,H generallydependsonthenumber
EB
two curves cross again, indicating a second field-induced
of M(H) measurements, a property often called train-
reconfiguration. This suggests the possible presence of ing effect1. The possible application of an EB material
twocontributionstotheZEBeffect,leadingtoadecrease
forspintronicdevicesrequiresittobe fairlystableunder
of the final magnetization which is probably related to
consecutive M(H) cycles. Since La Ca CoMnO ex-
1.5 0.5 6
the smaller ZEB effect observed for La Ca CoMnO
1.5 0.5 6 hibits both ZEB and CEB effects, we have investigated
in comparison to La Sr CoMnO .
1.5 0.5 6 the dynamics of the spin structure by measuring consec-
The EB field is here defined as HEB =(H++H−)/2, utive M(H) loops at 2 K in both ZFC and FC cases.
where H+ and H− represent the positive and negative For the CEB mode, the consecutive loops were car-
coercivefields,respectively. Theeffectivecoercivefieldis ried after cooling the system with H = 50 kOe. Fig.
HC =(H+−H−)/2. Fig. 6 shows that a negative ZEB 8 shows the first and sixth loops, wFhCere can be noted
effect is observed for La1.5Ca0.5CoMnO6 for T < 15 K, that the differences in the HCEB fields are mainly due
in resemblance to its analogue La1.5Sr0.5CoMnO69. But to variations in H−. On the upper inset of Fig. 8 it be-
here the magnitude of the ZFC EB field, HZEB, is much cames clearer that the magnitude of H− monotonically
smaller than for the Sr-based compound, which presents decreaseswiththenumberofconsecutivecycles(n),indi-
the largest ZEB effect reported so far. At T = 2 K one catingspinrearrangementatthe interfaces. The bottom
have HZEB ∼ 250 Oe for Ca-based sample, while for insetofFig. 8showstheevolutionofHCEB withn. This
Sr-based compound HZEB ∼ 6.5 kOe. Also interesting curvecanbefittoamodelconsideringtwocontributions
to note is the fact that initially HZEB increases with T. at the magnetic interfaces,the uncompensated rotatable
This resultis similartothatobservedforthe perovskite- spins and also the frozen spins10,43
based ZEB nanocomposite BiFeO -Bi Fe O 8. Possibly,
3 2 4 9
the small gain of thermal energy at low-T enables the |Hn |=|H∞ |+A e(−n/Pf)+A e(−n/Pr), (4)
CEB CEB f r
rotation of some AFM/CG spins to the field direction,
leading to the pinning ofthese spins after the removalof wheref andr denote the frozenandrotatablespincom-
magnetic field. ponents respectively. The fitting with Eq. 4 yields
∞
In order to further investigate the EB effect on H ∼ 2238 Oe, A ∼ 951 Oe, P ∼ 1.56, A ∼ 100
CEB f f r
La Ca CoMnO , we have also carried M(H) curves Oe, P ∼ 0.04. The precise fitting obtained when the
1.5 0.5 6 r
aftercoolingthe sampleinthepresenceofafieldH = glassyphaseistakenintoaccount,togetherwiththefact
FC
50 kOe. The inset of Fig. 7(a) shows the curve obtained thatallZEBmaterialsreportedsofarpresentsa SG-like
at 2 K, for which was observed an EB field H = phaseconcomitantlytoanothermagneticphase,indicate
CEB
7
(a)

3n = 1 /f.u.)B 100
2 n = 6 -3 0
u.) 1 M (1 16stht lloooopp e) 0 -9500 H90 (0O0e) 9300
/f.B 1 Hy2stere3sis n4umbe5r (n)6 (O 0.3
M (--021 -1.25H ( T)-1.20 H (kOe) CEB22..5705 H FCitE wBith Eq. 4 HZEB--210000 -3m()/f.u.M10B0.0 1st loop
2.25 -0.3 3rd loop
-3
-6 -4 -2 0 2 4 6 -300 9.26
H (T) (b)
9.6
9.24
FIG. 8: First and sixth loops obtained at 2 K, after cooling
the system with HFC = 50 kOe. The upper inset shows a )
magnifiedviewofH− forsixconsecutivecurves. Thebottom e) 9.4 Oe
insetshows-HCEB asafunctionofthehystereisnumber(n). kO |H-| 9.22(k
The solid line represents the fitting of experimental data to ( +
Eq. 4. -H 9.2 H+ H
-
9.20
the importance of the glassy phase to the observation of
9.0
EB effect on this and related PD compounds.
For the ZFC mode, an unprecedented behavior was 9.18
1 2 3 4
observed. As Fig. 9(a) shows, from the first to the sec-
Hysteresis number (n)
ond loop, H inverts its sign from negative to posi-
ZEB
tive. This result indicates that the ZFC procedure does
not ensure a good stability in the coupling between the
spinsresponsiblefortheEBeffect. Fromthesecondloop FIG.9: (a)HZEB asafunctionofthehysteresisnumber(n).
to subsequent measures,the system maintains a roughly Theinsetshowsamagnifiedviewofthefirstandthirdcurves
constant positive value. A closer inspection on the posi- close totheM =0axis. (b)H+ and-H− asafunction ofn,
obtainedaftertheZFCprotocol. Thelinesareguidesforthe
tive and negative coercive fields shows that the absolute
value of H− decreases for consecutive loops, while H+ eye.
increases [Fig. 9(b)].
Although this is the first report of sign inversion on
of H+ and the consequent inversion of H sign.
the EB field after the ZFC procedure, this behavior was ZEB
already observed for CEB systems. For NiFe/IrMn bi-
layer,the training-inducedinversionofH signalwas
CEB
IV. CONCLUSIONS
discussedin terms ofthe appearance ofmetastable mag-
netic disorder at the magnetically frustrated interface
during the hysteresis loop measurements43. Similarly to In summary, La Ca CoMnO is a RSG material
1.5 0.5 6
the SG model conjectured for EB systems in Refs.43,44, for which there are two anomalies on the magnetiza-
our results can be understood in terms of two contribu- tion as a function of T curve at T ≃ 158 K and
tions to the EB effect, the rotatable and frozen spins at T ≃138 K, probably related to Co2+–O–Mn4+ FM cou-
theAFM/FM/CGinterfaces. ForsuccessiveM(H)mea- pling and Co3+–O–Mn4+ AFM coupling. At T ≃ 46.7
g
surements,therearrangementoftherotatableAFM/FM K, a CG phase emerges. Similarly to the analogous
spinsattheinterfacewouldleadtotheexpecteddecrease La Sr CoMnO RSG compound, La Ca CoMnO
1.5 0.5 6 1.5 0.5 6
−
onthe magnitudeofH observedonFig. 9(b). Thesec- exhibit a non-negligible ZEB effect at low-T. But the
ondcontributionwouldarisefromtheCG/FMinterface. negative shift of the M(H) curve observed for the Ca-
Somespinsatthisinterfacewillprefertoalignantiparal- based material is substantially smaller than for Sr-based
lelto the directionofthe FM phase,definedby the posi- one. We believe that the lower symmetry of the Co–O–
tive field firstly applied in the first cycle. After training, Mn coupling for the Ca-doped sample and the particu-
these spins may rotate irreversibly due to the magneti- lar behavior of its initial magnetization observed at the
zation reversal of the FM phase, leading to the increase M(H) curve are responsible for such difference. When
8
the system is cooled in the presence of an applied mag- terials for application of the ZEB effect.
netic field, the EBeffect is greatlyenhanced. The fitting
of the training effect observed after the FC protocol in-
dicates the importance of the glassy phase to the EB
Acknowledgments
observedon this sample. For the ZFC procedure, H
ZEB
invertsits signfromnegativeto positivefromthefirstto
the second loop, which may be related to the unstable ThisworkwassupportedbyCNPq,CAPES,FAPESP
glassy spins at the interfaces. The presence of SG-like and FAPERJ (Brazil). L. Bufai¸cal thanks prof. S. A.
phase concomitantly to other magnetic phases seems to Le˜ao for the support. F. Stavale thanks the Surface and
be anecessaryconditiontotheobservationofZEB.This Nanostructures Multiuser Lab and the Lab of Biomate-
result should be taken into account when designing ma- rials at CBPF.
∗ Electronic address: [email protected] 20 B. Raveau and Md. Motin Seikh, Cobalt Oxides: From
1 J. Nogu´es and I. K. Schuller, J. Magn. Magn. Mater. 192 Crystal Chemistry to Physics (Wiley-VCH, Weinheim,
(1999) 203 . 2012).
2 J.Nogu´es,J.Sort,V.Langlais,V.Skumryev,S.Surin˜ach, 21 S.C.Petitto,E.M.Marsh,G.A.Carson,andM.A.Lan-
J. S.Mun˜oz, M. D.Baro´, Phys. Reports 422 (2005) 65. gell, J. Mol. Catal. A: Chem. 281 (2008) 49.
3 M. Ali, P. Adie, C. H. Marrows, D. Greig, B. Hickey, and 22 J. van Elp, J. L.Wieland, H.Eskes, P.Kuiper, and G. A.
R. L. Stamps,Nature Mater. 6(2007) 70. Sawatzky,Phys. Rev.B 44 (1991) 6090.
4 M. Gruyters,Phys. Rev.Lett. 95 (2005) 077204. 23 V. M. Jimenez, A. Fernandez, J. P. Espinos, and A. R.
5 E. Dagotto, Science 318 (2007) 1076. Gonzalez-Elipe, J.Electron.Spectrosc.Relat.Phenom.71
6 B. M. Wang, Y. Liu, P. Ren, B. Xia, K. B. Ruan, J. B. (1995) 61.
Yi, J. Ding, X. G. Li, and L. Wang, Phys. Rev. Lett. 106 24 R. Dudric, A. Vladescu, V. Rednic, M. Neumann, I.G.
(2011) 0077203 . Deac, and R.Tetean, J. Mol. Struct.1073 (2014) 66.
7 A. K. Nayak, M. Nicklas, S. Chadov, C. Shekhar, Y. Sk- 25 T. C. Koethe, Ph.D. thesis, Universit¨at zu K¨oln, 2007.
ourski, J. Winterlik, and C. Felser, Phys. Rev. Lett. 110 26 M. Hassel and H.-J. Freund, Surf. Sci. Spectra 4 (1996)
(2013) 127204. 273.
8 T.Maity,S.Goswami,D.Bhattacharya,andS.Roy,Phys. 27 C. A. F. Vaz, D. Prabhakaran, E. I. Altman, and V. E.
Rev.Lett. 110 (2013) 107201. Henrich,Phys. Rev.B 80 (2009) 155457.
9 J.KrishnaMurthyandA.Venimadhav,Appl.Phys.Lett. 28 C.R.Brundle,T.J.Chuang,andD.W.Rice,SurfaceSci.
103 (2013) 25410. 60 (1976) 286.
10 L. T. Coutrim, E. M. Bittar, F. Stavale, F. Garcia, E. 29 R.P. Gupta and S. K.Sen, Phys.Rev.B 12 (1975) 1.
Baggio-Saitovitch, M. Abbate, R. J. O. Mossanek, H. P. 30 M. C. Biesinger, B. P. Payne, A. P. Grosvenor, L. W. M.
Martins, D. Tobia, P. G. Pagliuso, and L. Bufai¸cal, Phys. Lau,A.R.Gerson,andR.St.C.Smart,Appl.SurfaceSci.
Rev.B 93 (2016) 174406. 257 (2011) 2717.
11 J.KrishnaMurthy,K.D.Chandrasekhar,H.C.Wu,H.D. 31 M.Oku,K.Hirokawa,andS.Ikeda,J.ElectronSpectrosc.
Yang, J. Y. Lin, and A. Venimadhav, J. Phys.: Condens. Relat. Phenom. 7 (1975) 465.
Matter 28 (2016) 086003. 32 V.diCastroandG.Polzonetti, J.Electron Spectrosc.Re-
12 N. Narayanan, D. Mikhailova, A. Senyshyn, D. M. Trots, lat. Phenom. 48 (1989) 117.
R. Laskowski, P. Blaha, K. Schwarz, H. Fuess, and H. 33 S. Ardizzone, C.L. Bianchi, and D. Tirelli, Colloids Surf.
Ehrenberg, Phys. Rev.B 82 (2010) 024403. A: Physicochem. Eng. Asp. 134 (1998) 305.
13 T. Fukushima, A. Stroppa, S. Picozzi, and J. M. Perez- 34 H.W.NesbittandD.Banerjee,Am.Mineralogist83(1998)
Mato, Phys. Chem. Chem. Phys. 13 (2011) 12186. 315.
14 R. D.Shannon,Acta Crystallographica A32 (1976) 751. 35 E.S.Ilton,J.E.Post,P.J.Heaney,F.T.Ling,S.N.Kerisit,
15 A.C.Larson,R.B.VonDreele,GeneralStructureAnalysis Appl.Surf. Sci. 366 (2016) 475.
System GSAS, Los Alamos National Laboratory Report 36 D.Niebieskikwiat, R.D.S´anchez,A.Caneiro, L.Morales,
(2001). M.V´asquez-Mansilla,F.RivadullaandL.E.Hueso,Phys.
16 J. Krishna Murthy, K. Devi Chandrasekhar, S. Muru- Rev.B 62 (2000) 3340.
gavelc, and A. Venimadhav, J. Mater. Chem C 3 (2015) 37 N. W. Ashcroft and N. D. Mermin, Solid State Physics,
836. Cengage Learning (1976).
17 H. Gamari-Seale, I. O. Troyanchuk, A. P. Sazonov, K. L. 38 R.L. Carlin, Magnetochemistry, Springer, Berlin (1986).
Stefanopoulos,andD.M.Toebbens,PhysicaB403(2008) 39 J.B.Goodenough,Magnetism and Chemical Bond (Inter-
2924. science, New York,1963).
18 J. Androulakis, N. Katsarakis, J. Giapintzakis, N. 40 J.A.Mydosh,SpinGlasses: AnExperimentalIntroduction
Vouroutzis, E. Pavlidou, K. Chrissafis, E. K. Polychro- (Taylor & Francis, London, 1993).
niadis,andV.Perdikatsis,J.J.SolidStateChemistry173 41 V.K.Anand,D.T. Adroja, andA. D.Hillier, Phys. Rev.
(2003) 350. B 85 (2012) 014418.
19 D. Serrate, J. M. De Teresa, and M. R. Ibarra, J. Phys.: 42 A.Malinowski,V.L.Bezusyy,R.Minikayev,P.Dziawa,Y.
Comdens. Matter 19 (2007) 023201. Syryanyy,andM.Sawicki,Phys.Rev.B84(2011)024409.
9
43 S. K. Mishra, F. Radu, H. A. Durr, and W. Eberhardt, (2008) 97.
Phys. Rev.Lett. 102 (2009) 177208.
44 F. Radu and H. Zabel, Springer Tracts Mod. Phys. 227