Canted Magnetic Ground State of Quarter-Doped Manganites $R_{0.75}$Ca$_{0.25}$MnO$_3$ ($R$ = Y, Tb, Dy, Ho, and Er)

1.7 MB·

by  R. Sinclair

Canted Magnetic Ground State of Quarter-Doped
Manganites R Ca MnO (R = Y, Tb, Dy, Ho, and Er)
0.75 0.25 3
R. Sinclair,1 H. B. Cao,2 V. O. Garlea,2 M. Lee,3,4 E. S. Choi,4
Z. L. Dun,1 S. Dong,5,∗ E. Dagotto,1,6 and H. D. Zhou1,4,†
1Department of Physics and Astronomy, University of Tennessee, Knoxville, Tennessee 37996-1200, USA
2Quantum Condensed Matter Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee, 37831, USA
3Department of Physics, Florida State University, Tallahassee, Florida 32306, USA
4National High Magnetic Field Laboratory, Florida State University, Tallahassee, Florida 32310, USA
5Department of Physics, Southeast University, Nanjing 211189, China
6Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
6 (Dated: January 12, 2016)
1
0 Polycrystalline samples of the quarter-doped manganites R0.75Ca0.25MnO3 (R = Y, Tb, Dy,
2 Ho, and Er) were studied by X-ray diffraction and AC/DC susceptibility measurements. All five
samplesareorthorhombicandexhibitsimilarmagneticproperties: enhancedferromagnetismbelow
n T (∼80 K) and a spin glass (SG) state below T (∼30 K). With increasing R3+ ionic size, both
a 1 SG
T and T generally increase. The single crystal neutron diffraction results on Tb Ca MnO
J 1 SG 0.75 0.25 3
revealedthattheSGstateismainlycomposedofashort-rangeorderedversionofanovelcanted(i.e.
0
noncollinear)antiferromagneticspinstate. Furthermore,calculationsbasedonthedoubleexchange
1
modelforquarter-dopedmanganitesrevealthatthisnewmagneticphaseprovidesatransitionstate

between the ferromagnetic state and the theoretically predicted spin-orthogonal stripe phase.
]
i
c
PACSnumbers: 72.80.Ga,71.30.+h,75.50.Dd,61.05.cp
s
-
l
r
t I. INTRODUCTION teraction between the t2g spins, JAFM, is small in agree-
m ment with the phase diagram of large and intermediate
t. Recentresearchinthefieldofmanganiteshasprimarily bandwidth manganites. However, a new exotic multi-
a focusedonthemultiferroicpropertiesofRMnO .1–9This ferroic phase, dubbed the spin-orthogonal-stripe (SOS)
m 3 phase, was found at the large J end of the narrow-
includes studies of Type-I multiferroics employing mate- AFM
- rialssuchasthehexagonalYMnO whereferroelectricity bandwidth manganites. The SOS state, as shown in Fig.
d and magnetism have different ori3gins,10 and also stud- 2(a) of Ref. 20, is made of zigzag chains, as in the CE-
n
phase of manganites at quarter doping, but forming an
o ies of Type-II multiferroics as in the cases of TbMnO3
array of diagonally oriented domains with spins rotated
c and orthorhombic HoMnO where the ferroelectricity is
[ causedbypeculiarmagnetic3orders.11–13Allofthesemul- by 90° between domains. Generalizations to other hole
tiferroic RMnO materials are located in the narrow- dopings have been studied as well.21
1 3
v bandwidth limit of manganites due to their small R3+ The intriguing prediction of an SOS state re-
3 ionic size.14 The fact that in Type-II multiferroics the quires detailed experimental studies. Although
7 ferroelectricity is strongly coupled to magnetism makes a couple of previous efforts have been reported
2 it very exciting to explore the possibility of new mag- for the doped manganites with small R3+ ions,
2 netic states in narrow-bandwidth systems. Discovering such as Tb1−xCaxMnO3,22–27 Dy1−xCaxMnO3,28 and
.0 exotic magnetic phases may lead to functional multifer- Y1−xCaxMnO3,29,30 asystematicexplorationofquarter-
1 roics with high transition temperatures and large spon- doped manganites R0.75Ca0.25MnO3 has not been con-
0 taneous polarizations. ducted and, for this reason, the existence of the SOS
6
Whilethecasesofpureundopedmanganiteshavebeen phase,orotherexoticmagneticphases,hasnotbeencon-
1
: extensively studied,15,16 the recent exploration for po- firmed so far. Moreover, theoretically the nature of the
v tential new magnetic phases has focused on doped man- quarter-doped manganites with an intermediate JAFM
Xi ganitesinthenarrow-bandwidthlimit.17,18 Thecompeti- strength has not been studied in full detail: it is only
tion between the ferromagnetic double exchange interac- knownthattheFMandSOSstatesarestableinthelim-
r
a tions and the antiferromagnetic superexchange interac- its of small and large JAFM, respectively.20 Thus, an in-
tions, plus the robust Jahn-Teller spin-lattice coupling, triguingquestiondevelops: doesanynewmagneticphase
makesthedopednarrow-bandwidthmanganitesanideal exist in the intermediate coupling region in between the
playground to search for new magnetic phases.19 More- FM and the predicted SOS phases? Plenty of previous
over, considerable theoretical progress has been made in theoreticalstudieshaveconsistentlyshownthatmangan-
this area of research. For example, the magnetic phase ites in general, i.e. not only multiferroics, have the po-
diagram of quarter-doped manganites has been investi- tential to display a wide variety of complex patterns of
gated based on the double exchange model.20 A promi- spin, charge, and orbital order.31–35
nent ferromagnetic (FM) metallic orbitally-disordered With this timely fresh motivation in mind to search
phase occupies the region where the superexchange in- for possible new magnetic phases of quarter-doped man-
2
ganites with a narrow-bandwidth, in this publication III. RESULTS
we have systematically studied the magnetic character-
istics of R Ca MnO (R = Tb, Dy, Ho, Y, and
0.75 0.25 3 A. Polycrystalline R0.75Ca0.25MnO3
Er). First, we analyzed the magnetic properties of all
five polycrystalline samples and found that their general
Figures 1 (a-e) display the room temperature XRD
behavior consists of enhanced FM interactions below 80
patterns for R Ca MnO (R = Tb, Dy, Ho, Y, and
K and a spin glass state below 30 K. Second, we per- 0.75 0.25 3
Er). All samples show a pure orthorhombic (Pbnm)
formedneutronscatteringexperimentsonasinglecrystal
structure. With decreasing ionic size R, the lattice
of Tb Ca MnO and discovered that its magnetic
0.75 0.25 3 parameters decrease, as shown in Fig. 1(f). The at-
ground state is dominated by the short range ordering
tempts to prepare orthorhombic R Ca MnO with
of a new canted spin state which results in the observed 0.75 0.25 3
R3+ ions smaller than Er3+ failed. Noting that pure
spin glass-like behavior. Finally, we conducted theoreti-
HoMnO , YMnO , and ErMnO have hexagonal struc-
calcalculationsbasedonthedoubleexchangemodelcon- 3 3 3
tures, it seems that the substitution of Ca2+ can stabi-
firming that the observed new spin structure is lower in
lize the perovskite structure. This is easy to understand
energythanboththeFMandSOSstatesintheinterme-
since Ca2+ ions are larger than the R3+ ions involved
diate strength region of J ; in other words, the novel
AFM here and this doping will reduce the manganese size.
magnetic phase reported here acts as a transition state
Therefore, the structural tolerance factor (t) will be in-
between the FM and the SOS phases.
creasedand,generally,astincreasesthedrivingforcefor
the octahedral rotation increases as well, leading to the
transformation from the hexagonal to the orthorhombic
phase. Several reported studies on Ho Ca MnO and
1−x x 3
Y Ca MnO have confirmed this observation.17,30,37
1−x x 3
Figure 2 shows the DC susceptibility (χ) results
II. EXPERIMENTAL DETAILS corresponding to R0.75Ca0.25MnO3. For the case of
Tb Ca MnO , χ shows two major features: (i) a
0.75 0.25 3
sharp increase around 80 K with decreasing tempera-
Polycrystalline samples of R Ca MnO (R = ture. This feature is more evident as a slope change
0.75 0.25 3
Tb, Dy, Ho, Y, and Er) were synthesized by from the linear temperature dependence of the inverse
solid state reactions. The stoichiometric mixture of the susceptibility (Fig. 2(f)). Here we define the
of Tb O /Dy O /Ho O /Y O /Er O , CaCO , and peak position of the derivative of 1/χ as T . Clearly,
4 7 2 3 2 3 2 3 2 3 3 1
Mn O were ground together and then calcined in air ferromagnetic tendencies develop below T ; (ii) a broad
2 3 1
at 950 ◦C, 1200 ◦C, and 1350 ◦C for 24 hours, respec- peak around 34 K where zero field cooling (ZFC) and
tively. Single crystals of Tb Ca MnO were grown field cooling (FC) curves display a large splitting. This
0.75 0.25 3
by the traveling-solvent floating-zone (TSFZ) technique peakpositionisdefinedasT . Similarly, allothersam-
SG
inanIR-heatedimagefurnace(NEC)equippedwithtwo ples also present these two features as well. As shown
halogen lamps and double ellipsoidal mirrors. The crys- in Fig. 3, with increasing R3+ ionic size, both T and
1
tal growth rate was 15 mm/h. Small pieces of single T generallyincrease. Oneimportantpointhereisthat
SG
crystalsweregroundintoafinepowderforX-raydiffrac- since Y Ca MnO with nonmagnetic Y3+ ions dis-
0.75 0.25 3
tion. TheresultingpowderX-raydiffraction(XRD)pat- playthesetwofeatures,theymustberelatedtothemag-
ternswererecordedatroomtemperaturewithaHUBER netic Mn3+/Mn4+ ions. To further investigate the mag-
ImagingPlateGuinierCamera670withGemonochrom- netic properties, we performed a Curie-Weiss fit on the
atizedCuK radiation(1.54059˚A).Thelatticeparam- 1/χ data, as shown in Fig. 2(i), resulting in a Curie
α1
eters were refined from the XRD patterns by using the temperature of θ = 58.1 K and an effective magnetic
CW
software package FullProf Suite with typical refinements momentofµ =5.16 µ . ForY Ca MnO ,there
eff B 0.75 0.25 3
for all samples having χ2 ≈ 0.7. X-ray Laue diffraction are75%Mn3+ ions(µ ≈4.8 µ )and25%Mn4+ ions
eff B
was used to align the crystals. Elastic neutron scatter- (µ ≈ 3.8 µ ) in the system, resulting in an expected
eff B
ing measurements were performed at the Neutron Pow- totaleffectivemomentofµ ≈4.6 µ whichisconsis-
eff B
der Diffractometer (HB-2A), and single-crystal neutron tent with our crude fitting analysis.
scattering measurements were performed at the Four- The AC susceptibility was measured for
Circle Diffractometer (HB-3A). Both instruments are lo- Tb Ca MnO to study the nature of the tran-
0.75 0.25 3
cated at the High Flux Isotope Reactor (HFIR) in Oak sition at T . As shown in Fig. 4(a), around T
SG SG
Ridge National Laboratory (ORNL). The neutron scat- the AC susceptibility presents a frequency dependent
tering diffraction patterns were also refined using Full- peak. The Mydosh parameter ∆T /[T ∆log(f)], a
SG SG
Prof Suite. The DC magnetic-susceptibility measure- quantitative measure of the frequency shift, is estimated
ments were performed employing a Quantum Design su- to be 0.034 (Fig. 4(b)). This is of the same order as the
perconducting interference device (SQUID) magnetome- expected range of 0.004-0.018 for conventional spin glass
ter. TheACsusceptibilitydatawasmeasuredonahome- systems.38 The ZFC and FC splitting feature is also a
made setup.36 characteristic behavior of a spin glass transition.
3
FIG. 1: (color online) Room temperature XRD patterns for polycrystalline R Ca MnO : (a) Er, (b) Ho, (c) Tb, (d)
0.75 0.25 3
Y, and (e) Dy. The crosses are the experimental data. The solid curves are the best fits from the Rietveld refinement using
FullProfSuite. TheverticalmarksindicatethepositionoftheBraggpeaks,andthebottomcurvesshowthedifferencebetween
the observed and calculated intensities. (f) The R3+ ionic size dependence of the lattice parameters.
The powder neutron diffraction measurements were at Q = 1.2 ˚A−1, matching the (100) peak position. The
performed on Tb Ca MnO , and the differential presence of this lattice forbidden (100) peak should be
0.75 0.25 3
scattering results are shown in Fig. 5. The large fea- associated with the development of an antiferromagnetic
ture near Q = 2.8 ˚A−1 is due to the change in lattice ordering. It is noteworthy that even at 4 K, these peaks
parameters. The negative differential signal at Q < 0.5 are still broader than the instrument resolution, and the
˚A−1 iscausedbythereductionintheparamagneticscat- correlation length derived from the (100) Lorentzian full
tering that exists above the ordering temperature. Note peakwidthathalfmaximumisapproximatelyξ =50˚A.
that the paramagnetic scattering follows the Q depen-
dence of the Mn3+/Mn4+ magnetic form factor. The
observed magnetic peak at Q=1.6 ˚A−1 is broader than
the instrument resolution. Its position suggests it is due
to scattering from a short range ferromagnetic ordering
thatcontributestothe(110)or(002)Braggpeakswhich Therefore, it is concluded that Tb Ca MnO de-
0.75 0.25 3
can be explained by a ferromagnetic ordering along ei- velops ferromagnetic characteristics below T and enters
1
ther the a or b axes. This is consistent with the DC sus- a short-range ordered state below T , but the latter
SG
ceptibility results showing the development of ferromag- containsantiferromagnetictendencies. Basedonthesim-
netism below T1. Figure 5(b) contains the differential ilarity among the DC susceptibility measurements of all
scatteringbetween4Kand120K.Thedatashowsmore the R Ca MnO samples, this development of mag-
0.75 0.25 3
pronounced scattering that adds to the (110) or (002) netism in the Tb sample should occur in all the other
peaks. In addition, there is another new magnetic peak samples as well.
4
FIG. 2: (color online) Temperature dependence of the DC susceptibility for polycrystalline R Ca MnO : (a) Tb, (b) Dy,
0.75 0.25 3
(c)Ho,(d)Y,and(e)Er. Temperaturedependenceoftheinverseofthesusceptibilityanditsderivativefor(f)Tb,(g)Dy,(h)
Ho, (i) Y, and (j) Er. In (i), the linear solid line represents the Curie-Weiss fit for the T > 150 K data.
B. Single crystal Tb Ca MnO single crystal is presented in Fig. 6.
0.75 0.25 3
Figure 7 shows the DC susceptibility of the
Tb Ca MnO single crystal with an applied mag-
0.75 0.25 3
In order to further clarify the nature of the magnetic neticfieldH alongdifferentaxes. ForH||aandH||c,the
properties of R Ca MnO , we tried to grow single results for χ are similar between the two data sets which
0.75 0.25 3
crystals for more detailed studies. We successfully grew showsaslopechangearound80K,asdefinedbythepeak
single crystals of Tb Ca MnO by using the float- obtained from the derivative of 1/χ, and a ZFC and FC
0.75 0.25 3
ing zone technique. The obtained crystals cleave easily splittingaround30K.Thesefeaturesareconsistentwith
into several millimeter-long needle pieces. The attempts thepolycrystallineresults. ForH||b,theχresultsaredif-
to grow other R Ca MnO samples all failed. No- ferent from those corresponding to H||a and H||c. First,
0.75 0.25 3
tably, after melting at high temperatures, the Ho, Y, the sharp increase of χ now shifts to around 50 K. Sec-
andErsamplesshowphaseseparationbyintroducingthe ond, theFCcurvekeepsincreasingbelow30Kwhilethe
hexagonal phase into the orthorhombic phase. This fact ZFCcurveactuallyshowsnegativevalueswhichstrongly
suggests that the orthorhombic phase of the Ho, Y, and suggests the existence of ferromagnetic domains.
Ersamplesisameta-stablephaseatlowtemperatures. A As shown in Fig. 8, the single crystal neutron diffrac-
Laue pattern measured on the grown Tb Ca MnO tion measurements on Tb Ca MnO display two
0.75 0.25 3 0.75 0.25 3
5
TABLE I: Structural parameters for the R Ca MnO samples (R = Y, Tb, Dy, Ho, and Er) at room temperature (space
0.75 0.25 3
group Pbnm) determined from refined XRD measurements.
Refinement Atom Site x y z Occupancy
Y 4c -0.01519(29) 0.07107(15) 1/4 0.37365(97)
XRD
Ca 4c -0.01519(29) 0.07107(15) 1/4 0.12635(97)
R = Y
Mn 4b 1/2 0 0 0.50
χ2 = 0.976
O1 4c 0.11811(86) 0.45828(96) 1/4 0.50
(a)
O2 8d 0.69571(76) 0.29496(73) 0.04660(71) 1.00
a = 5.290982(47), b = 5.626886(54), c = 7.448109(62)
Overall B-factor = 1.9561
Tb 4c -0.01760(61) 0.06673(33) 1/4 0.37403(133)
XRD
Ca 4c -0.01760(61) 0.06673(33) 1/4 0.12597(133)
R = Tb
Mn 4b 1/2 0 0 0.50
χ2 = 0.332
O1 4c 0.09883(233) 0.46518(266) 1/4 0.50
(b)
O2 8d 0.70392(241) 0.29343(216) 0.04061(225) 1.00
a = 5.333127(126), b = 5.628304(138), c = 7.493484(167)
Overall B-factor = 2.4558
Dy 4c -0.01432(49) 0.06895(24) 1/4 0.34640(94)
XRD
Ca 4c -0.01432(49) 0.06895(24) 1/4 0.15360(94)
R = Dy
Mn 4b 1/2 0 0 0.50
χ2 = 0.697
O1 4c 0.09971(160) 0.47354(181) 1/4 0.50
(c)
O2 8d 0.69111(145) 0.30316(136) 0.04818(147) 1.00
a = 5.317957(89), b = 5.624245(103), c = 7.482448(124)
Overall B-factor = 2.3016
Ho 4c -0.01513(31) 0.06859(16) 1/4 0.37244(69)
XRD
Ca 4c -0.01513(31) 0.06859(16) 1/4 0.12756(69)
R = Ho
Mn 4b 1/4 0 0 0.50
χ2 = 0.628
O1 4c 0.10984(114) 0.48981(127) 1/4 0.50
(d)
O2 8d 0.70930(115) 0.31137(87) 0.05731(90) 1.00
a = 5.291904(53), b = 5.640843(58), c = 7.449123(70)
Overall B-factor = 2.2878
Er 4c -0.01729(20) 0.07180(12) 1/4 0.37492(51)
XRD
Ca 4c -0.01729(20) 0.07180(12) 1/4 0.12508(51)
R = Er
Mn 4b 1/4 0 0 0.50
χ2 = 0.616
O1 4c 0.10613(96) 0.46861(103) 1/4 0.50
(e)
O2 8d 0.71026(92) 0.31032(74) 0.05153(64) 1.00
a = 5.269618(41), b = 5.647665(42), c = 7.428105(54)
Overall B-factor = 1.8764
magnetic signals at low temperatures: (i) a weak broad magnetic ordering for temperatures lower than 35 K.
peakaroundtheantiferromagneticBraggposition(102);
The details of the spin structure are shown in Fig. 9.
(ii) a strong broad peak around the ferromagnetic Bragg
Here, we see that the Mn spins are canted both in and
position (002). The temperature dependence of the in- out of the a-b plane with canting angles of ∼77◦ and
tensity of the (102) peak shows that it develops below ∼32◦, respectively. The total refined moment of the sys-
35 K, which is around T . Meanwhile, the intensity of
SG tem was 3.689 µ , as shown in Table II, smaller than
the shoulder of the (002) peak at θ = 6.5° starts devel- B
our crude Curie-Weiss fit but still robust. This spin
opingbelow80K(T )andsharplyincreasesbelow35K.
1 structure is different from the canonical A-type, E-type,
These features are consistent with the powder neutron
CE-type, C-type, G-type, and spiral-type antiferromag-
diffraction results. The refinement of this single crystal
netic phases observed in manganites before (using the
neutron diffraction data leads to the identification of the canonicalnotation39,40). Althoughthesinglecrystalneu-
6
FIG. 3: (color online) Variations of the T and T tem-
1 SG
peratures with the R3+ ionic size for the R Ca MnO
0.75 0.25 3
materials studied here.
trondataenablesustoresolvetheantiferromagneticspin
structure, the broadness of the magnetic Bragg peaks
stronglysuggeststhatthisantiferromagneticorderingei-
ther has short range characteristics or it has a clustered
nature, as opposed to a true long range magnetic order-
ing.
IV. DISCUSSION
Only a few previous studies of narrow bandwidth
R Ca MnO have addressed the magnetic properties
1−x x 3
of R Ca MnO .17,20 For example, Blasco et al. an-
0.75 0.25 3
alyzed Tb Ca MnO and they showed that with in- FIG. 4: (color online) (a) Temperature dependence of
1−x x 3
creasing Ca doping the ferromagnetic interactions are the real part of the AC susceptibility for polycrystalline
enhanced while the antiferromagnetic ordering is sup- Tb0.75Ca0.25MnO3 under different frequencies. (b) The fre-
quency dependence of ∆T /T .
pressed, and in particular the x = 0.25 sample has SG SG
a spin glass ground state.22 Pena et al. focused on
Dy Ca MnO and they also showed an enhanced fer-
1−x x 3
romagnetic interaction around 80 K for the x = 0.25 discussedbelowthespinglassgroundstateiscompatible
sample.28Inaddition,aneutronpowderdiffractionstudy withashortrangeorderingversionofanovelcantedfer-
of Y Ca MnO , which has a similar composition as romagneticspinstatethatwasunveiledbasedonourde-
0.7 0.3 3
the Y0.75Ca0.25MnO3 case studied here, shows that the tailed neutron diffraction studies on Tb0.75Ca0.25MnO3
magnetic ground state has short range ordering with (see Fig. 9).
an antiferromagnetic nature due to the observed broad To verify the exotic magnetic pattern obtained in
Bragg magnetic peak around the (001) reflection.30 Our our neutron analysis, here a microscopic theoretical
reported data here is consistent with all of these previ- study is performed based on the standard two-orbital
ous results. More importantly, our systematic studies of double-exchange model.39–41 In the past decade, this
R Ca MnO pointoutthepresenceofenhancedfer- modelHamiltonianhasbeenwidelyusedtoinvestigatea
0.75 0.25 3
romagnetictendenciesaround80K(T )andalsotheex- plethora of magnetic phases and their associated phys-
1
istence of spin glass behavior around 30 K (T ). More- ical characteristics, such as colossal magnetoresistance
SG
over,thesearegeneralbehaviorsforallR Ca MnO and multiferroicity, in perovskite manganites.39,40 The
0.75 0.25 3
withR=Tb,Dy,Ho,Y,andEr. WithincreasingRionic clearsuccessofthepreviouseffortsinthiscontextallows
size, both T and T generally increase; in addition, as us to investigate with confidence the possibility of new
1 SG
7
12 (a) 40 K - 120K
(110)
)
it 8 FM
n (002)
u FM
. 4
b
r
a
( 0
g
n
ri -4
e
tt 12 (b) (110)
a FM
sc (002) 4 K - 120K
e 8 (100) FM
c AFM
n
e 4
r
e
f
f
i 0
D
-4
1 2 3 4 5
Q (Å-1)
FIG.5: Thedifferentialscatteringpatternforpolycrystalline
Tb Ca MnO (a) between 40 K and 120 K and (b) be-
0.75 0.25 3
tween 4 K and 120 K.
FIG. 6: The Laue pattern oriented along the [001] direction
for the grown Tb Ca MnO single crystal.
0.75 0.25 3
phases in previously unexplored regions of the phase di-
agrams via the double exchange model. More explicitly,
the model Hamiltonian used here reads as:
αβ
(cid:88) (cid:88)
H =− tr (Ω c† c +h.c.)+J S ·S . FIG. 7: (color online) Temperature dependence of the DC
αβ ij i,α j,β AFM i j
susceptibility for a single crystal of Tb Ca MnO with
<ij> <ij> 0.75 0.25 3
(a) H||a, (b) H||b, and (c) H||c. (d) Derivative of 1/χ with
(1)
respect to temperature, with H along the three axes.
This Hamiltonian contains two terms. The first
term denotes the standard two-orbital double-exchange
hopping process for the e electrons between nearest-
g
8
250 (a) (1 0 2)
200 4 K
56 K
150
100 K

100
50
) 4 5 6 7 8 9 10 11 12 13 14 15
te q (degre es)
u
n
i
m1600 (0 0 2)
(b)
s/ 10 K
t1200
n 200 K
u
o 800
c
(
y 400
t
i
s
n 0
e
t 2 3 4 5 6 7 8 9 10 11 12 13
n
I q (deg rees)
(c) (102)
600
(002) shoulder at q = 6.5(cid:176)
400

200
0
0 20 40 60 80 100 120 140 160
T (K)
FIG. 9: The novel spin state reported for the
Tb Ca MnO lattice. The manganese sites Mn1,
0.75 0.25 3
FIG. 8: (color online) The (a) (102) and (b) (002) peaks at Mn2, Mn3, and Mn4 of Table II are indicated. There are
differenttemperaturesfortheTb0.75Ca0.25MnO3 singlecrys- two canting angles between Mn’s nearest-neighbor spins: an
tal. (c) Temperature dependence of the (102) peak intensity in-plane angle of ∼77◦ and an out-of-plane angle of ∼32◦.
and the (102) peak shoulder (θ = 6.5°) intensity.
hopping amplitudes along the three axes given by:
√
(cid:18)tx tx (cid:19) t (cid:18) 3 − 3(cid:19)
tx = aa ab = 0 √ ,
tx tx 4 − 3 1
ba bb
√
(cid:18)ty ty (cid:19) t (cid:18) 3 3(cid:19)
ty = aa ab = 0 √ ,
ty ty 4 3 1
ba bb
(cid:18)tz tz (cid:19) (cid:18)0 0(cid:19)
neighborsitesiandj ofathreedimensionalcubiclattice tz = taza tazb =t0 0 1 . (2)
for the manganese ions. The operators c† (c ) cre- ba bb
i,α j,β
ate (annihilate) an e electron at the orbital α (β) of In our calculations, the hopping amplitude t will be
g 0
the lattice site i (j). Working within the standard in- considered as the unit of energy. This hopping can be
finite Hund coupling approximation, shown to be qual- roughly estimated to be 0.5 eV.39,40 The second term of
itatively correct for manganites,39,40 the spin of the e the Hamiltonian is the antiferromagnetic superexchange
g
electronsisalwaysparalleltothespinofthelocalizedt interaction between the NN t spins.
2g 2g
degreesoffreedom,S,generatingtheBerryphase: Ω = The typical value of the superexchange coupling
ij
cos(θ /2)cos(θ /2)+sin(θ /2)sin(θ /2)exp[−i(φ −φ )], J in manganites is approximately 0.1t for the
i j i j i j AFM 0
where θ and φ are the polar and azimuthal angles of the morewidelystudiedmanganites,suchasLa Sr MnO
1−x x 3
classical t spins, respectively.39,40 The three nearest- and La Ca MnO , based on a variety of previous
2g 1−x x 3
neighbor(NN)hoppingdirectionsaredenotedbyr. Two investigations.39,40 The model studied here does not in-
e orbitals (a: x2 − y2 and b: 3z2 − r2) are involved clude the electron-lattice coupling, i.e. the Jahn-Teller
g
in the double-exchange process for manganites, with the distortions, but this coupling can be partially taken into
9
TABLE II: Magnetic moments for the single crystal Tb Ca MnO sample at 4.2 K (magnetic space group Pbn(cid:48)m(cid:48))
0.75 0.25 3
determined from refined neutron diffraction measurements.
x y z M (µ ) M (µ ) M (µ ) M (µ )
x B y B z B B
Mn1 1/2 0 0 2.886( 62) 2.063( 75) 1.011(153) 3.6888( 772)
Mn2 0 1/2 0 2.886( 62) -2.063( 75) -1.011(153) 3.6888( 772)
Mn3 1/2 0 1/2 2.886( 62) 2.063( 75) -1.011(153) 3.6888( 772)
Mn4 0 1/2 1/2 2.886( 62) -2.063( 75) 1.011(153) 3.6888( 772)
calculated in momentum space using a fine three di-
mensional grid and the results are shown in Fig. 10.
With increasing J , the ground state evolves from
AFM
the initial ferromagnetic state to the final SOS state, in
-0.85 New agreement with previous investigations.20 However, the
FM SOS most interesting new result is that in the middle region
A-AFM CE
t)0-0.90 C-AFM E-AFM (0.129t0 <JAFM <0.161t0),thenewlydiscoveredcanted
y ( spin order state displays the lowest energy among all of
g the candidates investigated here.
er -0.95
n According to these results, the case of
E
Tb Ca MnO should fall into this middle re-
0.75 0.25 3
-1.00
gion. The new canted phase provides a bridge between
the ferromagnetic and the SOS phases since it contains
-1.05 both a ferromagnetic and noncollinear components.
Thus, the pure SOS phase should be expected to be
0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18
stable only in even narrower bandwidth manganites.
JAFM It is worth clarifying that this canted state does
not induce ferroelectricity if we apply the inverse
Dzyaloshinskii-Moriya (DM) mechanism where lattice
FIG. 10: (color online) Energies (per site) of the magnetic
distortions leading to ferroelectricity are generated by
statesconsideredhere,asafunctionofJ . Theenergyunit
AFM specialnoncollinearmagneticstates.11Inthepresentcase
is t . FM: ferromagnetic; A-AFM: A-type antiferromagnetic;
0
the local displacements will compensate between nearest
C-AFM:C-typeantiferromagnetic;E-AFM:E-typeantiferro-
magnetic. These antiferromagnetic as well as the CE labels neighbors, and the global ferroelectric polarization will
are standard notations in manganites.39,40 The SOS state is cancel out. Although not ferroelectric, the noncoplanar
from Ref. 20. The novel canted state is denoted by “New”. spin texture of the state unveiled here is very novel, and
it can give rise to an intrinsic anomalous Hall effect.43,44
account by increasing the superexchange strength. Our
simplified model is expected to capture the main physics V. CONCLUSIONS
of manganites and, thus, can generate a phase diagram
qualitatively similar to the experimental observations.42 In this publication, we report detailed experimental
Therefore,itisacceptabletostudy,atleastqualitatively, studies of R Ca MnO (R = Y, Tb, Dy, Ho, and
0.75 0.25 3
thepropertiesofthemanganitesdiscussedinthepresent Er)polycrystals,andaTb Ca MnO singlecrystal,
0.75 0.25 3
publication using this simplified model. with focus on their magnetic properties. The amount of
For the quarter-doped manganites, previous investiga- Ca used corresponds to the hole quarter-doped case of
tions predicted a spin-orthogonal stripe SOS phase in the widely discussed manganite multiferroic perovskites.
the very narrow bandwidth region (J > 0.17t ).20 In general, we have observed the presence of ferromag-
AFM 0
The well known ferromagnetic phase observed in normal netic and spin-glass tendencies in all the samples stud-
manganitesappearsintheoppositeside(J <0.13t ). ied. Our main discovery, using the Tb-based single crys-
AFM 0
Inthemiddleregion,theMonteCarlosimulationdidnot tal, is that the spin-glass region appears dominated by
provide an unambiguous answer at that time because of the short-range order of a new canted, and thus non-
metastabilities in the Monte Carlo time evolution that collinear, magnetic state. The theoretical study of the
are typically indicative of complex magnetic patterns.20 double-exchange model presented here shows that in a
Sincethespinpatternobtainedintheneutronstudydis- reduced region of parameter space the new state has in-
cussed above is neither the SOS nor the normal ferro- deedlowerenergythanthetwostatespreviouslybelieved
magnetic phases, therefore it is necessary to recheck the to be dominant at quarter doping in narrow bandwidth
possible existence of new phases by taking into account manganites,namelytheFMandSOSstates. Theresults
the state discovered experimentally here. reported here illustrate that doped manganite multifer-
The energies for various fixed magnetic patterns are roic compounds harbor magnetic states that are more
10
complexthanpreviouslyanticipated. Ourpresentefforts HFIR/ORNL was sponsored by the Scientific User Fa-
areexpectedtopavethewayandmotivatemoredetailed cilities Division (H.B.C., O.V.G, J.M.), Office of Basic
studies of these mainly unexplored exotic materials that Energy Sciences, US Department of Energy. S.D. was
have potential for functional applications. supported by National Natural Science Foundation of
China (Grant Nos. 51322206). E.D. was supported by
theNationalScienceFoundationunderGrantNo. DMR-
Acknowledgments 1404375. The work at NHMFL is supported by Grant
No. NSF-DMR-1157490 and the State of Florida and by
R.S.,Z.L.D.,andH.D.Z.thankthesupportfromNSF- the additional funding from NHMFL User Collaboration
DMR through Award DMR-1350002. The research at Support Grant.
∗ Electronic address: [email protected] 24 N. Mufti, G. R. Blake, A. A. Nugroho, and T. T. M. Pal-
† Electronic address: [email protected] stra, J. Phys.: Condens. Matter 21, 452203 (2009).
1 T. Kimura, T. Goto, H. Shintani, K. Ishizaka, T. Arima, 25 A. Machida, Y. Moritomo, S. Mori, N. Yamamoto,
and Y. Tokura, Nature (London) 426, 55 (2003). K. Ohoyama, E. Nishibori, M. Takata, M. Sakata,
2 T. Goto, T. Kimura, G. Lawes, A. P. Ramirez, and Y. T. Otomo, and A. Nakamura, J. Phys. Soc. Jpn. 71, 27
Tokura, Phys. Rev. Lett. 92, 257201 (2004). (2002).
3 M. Kenzelmann, A. B. Harris, S. Jonas, C. Broholm, J. 26 A. Machida, Y. Moritomo, K. Ohoyama, and A. Naka-
Schefer,S.B.Kim,C.L.Zhang,S.-W.Cheong,O.P.Vajk, mura, J. Phys. Soc. Jpn. 70, 3739 (2001).
and J. W. Lynn, Phys. Rev. Lett. 95, 087206 (2005). 27 A. A. Nugroho, Risdiana, N. Mufti, T. T. M. Palstra,
4 T. Kimura, G. Lawes, T. Goto, Y. Tokura, and A. P. I. Watanabe, and M. O. Tjia, Physica B 404, 785 (2009).
Ramirez, Phys. Rev. B 71, 224425 (2005). 28 O. Pen˜a, M. Bahout, D. Gutierrez, P. Duran, and
5 T. Arima, A. Tokunaga, T. Goto, H. Kimura, Y. Noda, C. Moure, Solid State Sci. 5, 1217 (2003).
and Y. Tokura, Phys. Rev. Lett. 96, 097202 (2006). 29 J. R. Sahu, C. R. Serrao, A. Ghosh, A. Sundaresan, and
6 M. Mostovoy, Phys. Rev. Lett. 96, 067601 (2006). C. N. R. Rao, Solid State Commun. 149, 49 (2009).
7 B.Lorenz,Y.Q.Wang,andC.W.Chu,Phys.Rev.B76, 30 M. Dlouha´, S. Vratislav, Z. Jira´k, K. Kniˇzek, and D. Sed-
104405 (2007). midubsky´, Appl. Phys. A: Mater. Sci. Process. 74, S673
8 N. Hur, S. Park, P. A. Sharma, J. S. Ahn, S. Guha, and (2002).
S.-W. Cheong, Nature (London) 429, 392 (2004). 31 T.Hotta,S.Yunoki,M.Mayr,andE.Dagotto,Phys.Rev.
9 L. C. Chapon, G. R. Blake, M. J. Gutmann, S. Park, N. B 60, 15009 (1999).
Hur, P. G. Radaelli, and S.-W. Cheong, Phys. Rev. Lett. 32 T.Hotta,M.Moraghebi,A.Feiguin,A.Moreo,S.Yunoki,
93, 177402 (2004). and E. Dagotto, Phys. Rev. Lett. 90, 247203 (2003).
10 B.B.vanAken,T.T.M.Palstra,A.Filippetti,andN.A. 33 T.Hotta,A.Feiguin,andE.Dagotto,Phys.Rev.Lett.86,
Spaldin, Nature Mater. 3, 164 (2004). 4922 (2001).
11 I. A. Sergienko and E. Dagotto, Phys. Rev. B 73, 094434 34 T. Hotta, Y. Takada, H. Koizumi, and E. Dagotto, Phys.
(2006). Rev. Lett. 84, 2477 (2000).
12 I. A. Sergienko, C. S¸en, and E. Dagotto, Phys. Rev. Lett. 35 S. Dong, R. Yu, S. Yunoki, J.-M. Liu, and E. Dagotto,
97, 227204 (2006). Phys. Rev. B 78, 155121 (2008).
13 S. Picozzi, K. Yamauchi, B. Sanyal, I. A. Sergienko, and 36 Z. Dun, X. Li, R. S. Freitas, E. Arrighi, C. R. Dela Cruz,
E. Dagotto, Phys. Rev. Lett. 99, 227201 (2007). M. Lee, E. S. Choi, H. B. Cao, H. J. Silverstein, C.
14 A.Maignan,C.Martin,G.VanTendeloo,M.Hervieu,and R. Wiebe, J. G. Cheng, and H. D. Zhou, Phys. Rev. B
B. Raveau, Phys. Rev. B 60, 15214 (1999). 92, 024408 (2015).
15 S. Ishiwata, Y. Kaneko, Y. Tokunaga, Y. Taguchi, T. 37 P.P.RoutandB.K.Roul,J.Mater.Sci.: MaterElectron
Arima,andY.Tokura,Phys.Rev.B81,100411(R)(2010). 24, 2493 (2013).
16 S. Dong and J.-M. Liu, Mod. Phys. Lett. B 26, 1230004 38 J. A. Mydosh 1993 Spin Glasses: An Experimental Intro-
(2012). duction (London: Taylor and Francis).
17 N.Jiang,X.Zhang,andY.Yu,J.Phys.: Condens.Matter 39 E. Dagotto, Nanoscale Phase Separation and Colossal
25, 475901(2013). Magnetoresistance (Springer, Berlin, 2002).
18 K.Yoshii,H.Abe,andN.Ikeda,J.SolidStateChem.178, 40 E. Dagotto, T. Hotta, and A. Moreo, Phys. Rep. 344, 1
3615 (2005). (2001).
19 S.Demir,I.-R.Jeon,J.R.Long,andT.D.Harris,Coord. 41 E. Dagotto, J. Burgy, and A. Moreo, Solid State Commu-
Chem. Rev. 289-290, 149 (2015). nications 126, 9 (2003).
20 S. Dong, R. Yu, J.-M. Liu, and E. Dagotto, Phys. Rev. 42 S. Dong, X. T. Zhang, R. Yu, J.-M. Liu, and E. Dagotto,
Lett. 103, 107204 (2009). Phys. Rev. B 84, 155117 (2011).
21 S. Liang, M. Daghofer, S. Dong, C. S¸en, and E. Dagotto, 43 X.Chen,S.Dong,andJ.-M.Liu,Phys.Rev.B81,064420
Phys. Rev. B 84, 0024408 (2011). (2010).
22 J. Blasco, C. Ritter, J. Garc´ıa, J.M. de Teresa, J. P´erez- 44 N. Nagaosa, J. Sinova, S. Onoda, A. H. MacDonald, and
Cacho, and M.R. Ibarra, Phys. Rev. B 62, 5609 (2000). N. P. Ong, Rev. Mod. Phys. 82, 1539 (2010).
23 N. Mufti, A. A. Nugroho, G. R. Blake, and T. T. M. Pal-
stra, Phys. Rev. B 78, 024109 (2008).