Temperature dependence of iron local magnetic moment in phase-separated superconducting chalcogenide

0.76 MB·English

by  L. Simonelli

Temperature dependence of iron local magnetic moment in phase-separated
superconducting chalcogenide
L. Simonelli,1,2 T. Mizokawa,3,4 M. Moretti Sala,1 H. Takeya,5 Y.
Mizuguchi,3,5,6 Y. Takano,5 G. Garbarino,1 G. Monaco,1,7 and N.L. Saini3
1European Synchrotron Radiation Facility, BP220, F-38043 Grenoble Cedex, France
2CELLS - ALBA Synchrotron Radiation Facility, Carretera BP 1413,
Km 3.3 08290 Cerdanyola del Valles, Barcelona, Spain
3Dipartimento di Fisica, Universita´ di Roma “La Sapienza” - P. le Aldo Moro 2, 00185 Roma, Italy
4Department of Complexity Science and Engineering & Department of Physics,
University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8561, Japan
5
5National Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan
1
6Department of Electrical and Electronic Engineering, Tokyo Metropolitan University,
0
1-1 Minami-osawa, Hachioji, Tokyo, 192-0397, Japan
2
7Dipartimento di Fisica, Universita´ di Trento, I-38123 Povo, Trento, Italy

n (Dated: January 8, 2015)
a
J We have studied local magnetic moment and electronic phase separation in superconducting
7 KxFe2−ySe2 by x-ray emission and absorption spectroscopy. Detailed temperature dependent mea-
surements at the Fe K-edge have revealed coexisting electronic phases and their correlation with

] the transport properties. By cooling down, the local magnetic moment of Fe shows a sharp drop
n across the superconducting transition temperature (T ) and the coexisting phases exchange spec-
c
o tral weights with the low spin state gaining intensity at the expense of the higher spin state. After
c
annealing the sample across the iron-vacancy order temperature, the system does not recover the
-
r initial state and the spectral weight anomaly at Tc as well as superconductivity disappear. The
p results clearly underline that the coexistence of the low spin and high spin phases and the tran-
u sitions between them provide unusual magnetic fluctuations and have a fundamental role in the
s superconducting mechanism of electronically inhomogeneous K Fe Se system.
. x 2−y 2
t
a
PACSnumbers: 74.70.Xa,74.25.Jb,74.81.Bd
m
-
d
INTRODUCTION moment while ∼10-12% of the sample to remain para-
n
magnetic [17]. Similar conclusions have been reached
o
c by Nuclear magnetic resonance (NMR) experiments [16]
One of the interesting discoveries in the iron-based
[ showingdistinctspectraoriginatingfromamajorityanti-
chalcogenides has been a successful intercalation of FeSe
ferromagnetic and a minority metallic phase. The NMR
1
by alkaline atoms. The intercalated FeSe, forming a new
v results have also revealed a sharp drop in the relaxation
A Fe Se (A = K, Rb, Cs) family, shows a T as high
8 x 2−y 2 c rate and Knight shift on cooling across T , indicating
c
5 as 32 K [1–4] and, more importantly, exhibits several development of a field distribution due to the appear-
5 unique microstructural characteristics, such as a large
ance of a vortex state. Optical properties of supercon-
1 magnetic moment per Fe site, intrinsic Fe vacancy or-
ductingK Fe Se , haveshownthat∼10%samplevol-
0 x 2−y 2
derintheab-planeandanantiferromagneticorderinthe
. ume is covered by highly distorted metallic inclusions in
1 c-direction [5–7]. In addition, A Fe Se also shows
x 2−y 2 an insulating matrix, suggesting a filamentary network
0 phase separation [8–10], in which a magnetically or-
5 of conducting regions [18] and Josephson-coupling plas-
deredphasewithanexpandedlatticeandanonmagnetic
1 mons[19]. Inaddition,57Fe-Mo¨ssbauerspectroscopyhas
phase with a compressed lattice coexist [8]. The phase
: revealed phase separarion in ∼88% magnetic and ∼12%
v separation in A Fe Se has been studied by several
i x 2−y 2 nonmagnetic Fe2+ species [20]. A recent space resolved
X techniques [7–24], probing different properties. In par-
photoemission has provided further information on the
ticular, detailed temperature dependent neutron diffrac-
r peculiar electronic phase separation in metallic filamen-
a tion studies on superconducting A Fe Se (A=K, Rb,
x 2−y 2 tary phase and majority insulating texture [22]. Indeed,
Cs, Tl) have revealed same Fe vacancy ordered crystal
It is now known that the majority phase is insulating
structure and the same block antiferromagnetic order in
withastoichimetryofA Fe Se (245)containingiron
0.8 1.6 2
all these materials [7]. The intensity of the magnetic
vacancy order and block antiferromagnetism with large
Bragg peak, providing information on the iron magnetic
magnetic moment (∼3.3 µ ) while the minority phase is
B
moment, shows a continuous increase by lowering tem-
metallic with a stoichiometry of A Fe Se (122).
x 2 2
perature, however, this behavior is interrupted by a flat
plateau when T is approached. Muon spin relaxation Very recently, the coexisting electronic phases in
c
(µSR) experiments have shown phase separation with K Fe Se have been studied using high energy spec-
x 2−y 2
∼88-90%ofthesamplevolumetoexhibitlargemagnetic troscopy [24], revealing different phases to be character-
2
ized by different Fe valence and local magnetic moment.
In this work, we have further exploited x-ray emission K Fe Se K! (a)
0.03 x 2-y 2 1,3
(XES) and high resolution x-ray absorption (XAS) spec-
ttrivoistcyopayndtothsteudelyecctorrorneilca/tmioangbneettwiceepnhatsheerseudpisetrrciobnudtiuocn- nits) 0.03 K!
u 1,3 18K
as a function of temperature. While XES has been used b. 0.02 34K
toinvestigatetheevolutionoflocalFemagneticmoment y (ar 245000KK
µ,XAShasbeenexploitedtostudyunoccupiedelectronic sit 0.02 515K
states, corresponding to different electronic phases. The n 600K
magnetic moment µ shows anomalous temperature de- nte0.01 7054 7056
I
pendencewithasharpdropatthesuperconductingtran-
sition temperature. The results also reveal that the co- K!'
existing electronic phases with Fe2+ high spin (HS) and 0.00
7030 7040 7050 7060
low spin (LS) exchange their spectral weights across the
Energy (eV)
superconductingtransitiontemperature(∼32K).Onthe
otherhand,byannealingthesampleacrossthephasesep- KFe Se (b) (c) (d)
x 2-y 2
arationandiron-vacancyorderingtemperatures(∼520K FeSe
0.02
and ∼580 K respectively), the mean local moment µ is
)
decreased, and the HS-LS spectral exchange across Tc nits
gets suppressed with disappearance of the superconduc- u
tivity. rb.
(a x4 x4 x4
y 29 K 34 K 179 K
sit
EXPERIMENTAL DETAILS nten (e) (f) (g)
I
The XES and high resolution XAS measurements
were carried out on well characterized single crystals
of K Fe Se . After the growth using the Bridgman
x 2−y 2 0.00
method, the crystals were sealed into a quartz tube and x4 x4 x4
annealed for 12 hours in vaccum at 600◦C. The crys- 500 K 600 K 18 K Ann.
tals were characterized for their structure and phase pu- 7040 7060 7040 7060 7040 7060
rity by x-ray diffraction measurements [2]. The electric Energy (eV)
and magnetic characterizations were performed by resis-
tivity measurements in a PPMS (Quantum Design) and FIG. 1: (a) Fe Kβ XES measured on K Fe Se at sev-
x 2−y 2
magnetization measurements in a SQUID magnetometer eral temperatures. The spectra are normalized to the inte-
(Quantum Design). The samples exhibit a sharp super- grated area of the XES lines. The inset shows a zoom over
the main Kβ emission peak. (b)-(g) Fe Kβ emission spec-
conducting transition at T (cid:39) 32 K with a transition 1,3
c tra of K Fe Se (blue) at several temperatures are shown
width of ∼1 K in the temperature dependent resistivity x 2−y 2
compared with the one measured on FeSe at RT (red). The
signal. The superconducting volume fraction was found
differencespectra(green)arealsoshown(multipliedbyfour).
to be ∼10% at 2 K. The samples were exposed to air
only for a couple of minutes while mounting on the sam-
ple holder. For the rest of the time the samples were
manipulated in a glovbox or were kept in vacuum.
Thespectroscopymeasurementswereperformedatthe
beamline ID16 of the European Synchrotron Radiation
Facility. The experimental setup consists of a spectrom-
eter based on the simultaneous use of a bent analyzer low temperature measurements the sample was placed
Ge(620) crystal (bending radius R = 1 m) and a pixe- in a cryogenic environment, while for the high temper-
latedposition-sensitiveTimepixdetector[25]inRowland ature measurements it was placed in an oven. Firstly,
circlegeometry. Thescatteringplanewashorizontaland wehavecooleddownthesamplefromroomtemperature
paralleltothelinearpolarizationvectoroftheincidentx- (RT∼300 K) to 18 K, across T . This was followed by
c
rays. Themeasurementswerecarriedoutfixingthesam- warming the sample up to 600 K, destroying the phase
ple surface (the ab-plane of the K Fe Se single crys- separation (at ∼520 K) and disordering the Fe vacancies
x 2−y 2
tal) at ∼ 45◦ from the incoming beam direction and the (at ∼580 K). At the end, we have cooled down the sam-
scatteringangle2θ at∼90◦. Thetotalenergyresolution ple again to 18 K. The sample temperatures during the
was about 1.1 eV full width at half maximum. For the measurementswerecontrolledwithanaccuracyof±1K.
3
apparent changes across T [Fig. 1(b) and 1(c)]. A sub-
c
stantial change can be seen [Fig. 1(e) and 1(f)] across
the temperature for iron vacancy order-disorder (∼580
0.08 K). These variations are suppressed while the sample is
)
nits cooled after the annealing above the iron vacancy disor-
u der temperature [Fig. 1(g)].
rb. Cooling from RT Detailed temperature evolution of the IAD is shown
a 0.06 Heating from RT
D ( Cooling after anealing in Figure 2. By cooling from RT, the IAD shows an
A anomalousbehavior,representinganabnormalevolution
I
of µ. Indeed, the IAD shows a gradual decrease and
T
0.04 c an increase (with a local cusp-like anomaly at ∼ 170 K)
before showing a sharp drop at T . This is a clear indi-
c
0 100 200 300 400 500 600 cation of a lowering of the local magnetic moment in the
superconducting state. Earlier neutron diffraction stud-
Temperature (K)
ies on superconducting A Fe Se have shown that the
x 2−y 2
temperature dependent magnetic Bragg peak intensity
FIG. 2: Integrated absolute difference (IAD) calculated with
exhibits a plateau below T [7], indicating an interesting
respect to the FeSe as a function of temperature (see text). c
The thick continuous line (blue) is a guide to the eyes. The interplay between the ordered magnetic moment and su-
size of the symbols represents maximum uncertainty, esti- perconductivity. In the present work, the local magnetic
matedbyanalyzingdifferentXESscansateachtemperature. momentprobedbytheXESisfoundtoshowasharpdrop
atT ,revealingthatthereisHS-LStransitionandthere-
c
sulting magnetic fluctuations have a direct relation with
RESULTS AND DISCUSSIONS the superconductivity in phase separated K Fe Se .
x 2−y 2
TheIADshowsagradualdecreasewithincreasingtem-
Figure 1 (a) shows Fe Kβ emission spectra measured perature, revealing a transition-like behavior around the
onK Fe Se atdifferenttemperatures. Inthecrystal- phaseseparationandiron-vacancydisordertemperatures
x 2−y 2
field picture the overall spectral shape is dominated by (∼520 K and 580 K respectively). Indeed, by increasing
the (3p,3d) exchange interactions [26], and the presence temperature from RT to 600 K the Kβ emission line
1,3
(absence) of a pronounced feature at lower energy (Kβ’) shiftstowardslowerenergy[Fig. 1(a)]andtheKβ(cid:48) spec-
is an indication of a HS (LS) state of Fe [26]. The en- tral weight is suppressed. These results suggest that the
ergy position of the Kβ provides similar information system is undergoing a change in the spin configuration
1,3
on the spin state reflecting the effective number of un- across the phase separation temperature of ∼520 K [8],
paired 3d electrons [26]. A direct comparison between reaching a minimum around 580 K [Fig. 2] where the
measured spectrum on FeSe with those on K Fe Se blockantiferromagneticorderisknowntogetsuppressed.
x 2−y 2
(see, e.g. Fig. 1(b)to1(g))showsthatthelocalFemag- CoolingdowntoRT,thesampledoesnotrecovertheini-
netic moment µ is higher for the latter. It should be tial conditions, consistent with a glassy nature [30]. By
noted that the µ for the present sample is smaller than cooling after the annealing, the discontinuity across T ,
c
theoneforanair-exposedK Fe Se sample[24]. This evident in the cooling cycle, does not appear. Magnetic
x 2−y 2
furtherunderlinesfragilenatureofK Fe Se morphol- susceptibilitymeasurementsontheannealedsampledoes
x 2−y 2
ogy with physical properties being very sensitive to the not reveal any significant superconducting signal. This
air exposure. A zoom over the main Kβ emission off indicates that, either the annealed sample is no more su-
1,3
the K Fe Se is shown as inset of Fig. 1(a). Small perconducting or the superconducting volume fraction is
x 2−y 2
differences are apparent as a function of temperature. significantly reduced.
These variations are due to changing local Fe magnetic An independent and complementary information can
moment µ [26–29]. beobtainedfromXAS.Figure3showstheFeK-edgepar-
Integratedabsolutedifference(IAD)oftheKβ spectra tial fluorescence yield (PFY) XAS spectra measured on
with respect to a non magnetic reference is proportional K Fe Se at different temperatures. The spectra are
x 2−y 2
tothelocalµ[29]. Here,wehavedeterminedIADrelative measured by collecting the Fe Kβ emission as a func-
1,3
tothespectrumofFeSe. TheFeSeisnotanon-magnetic tion of incident energy, and normalized with respect to
reference,however,theIADstillprovidesusefulinforma- the atomic absorption estimated by a linear fit far away
tion on the relative evolution of µ as a function of tem- from the absorption edge. The XAS spectra show an
perature. Figure1(b)-1(g)showsFeKβemissionspectra asymmetric single pre-edge peak P [24, 31, 32], a shoul-
of K Fe Se at different temperatures compared with der A and broad near edge features B and C. The broad
x 2−y 2
the one measured on FeSe. The difference spectra pro- features indicate the system to have large atomic dis-
vides information of the changing local moment. The order, consistent with local structural studies [30]. The
differencespectraevolvewithtemperatureshowingsome mainpeakBdependsontemperatureindicatingachange
4
C (a)
1.0 B ) 0.80
2
) P
s +
nit 0.8 P1 0.78
b. u KFe Se P1/
(ar 0.6 x 2-y 2 ht ( 0.76
ty A 18 K eig
si 0.4 34 K w
Inten 0.2 P 250 K ative 0.74 T Cooling from RT
el c Heating from RT
R 0.72 Cooling after annealing
7110 7120 7130 7140 0 100 200 300 400 500 600
Energy (eV) Temperature (K)
0.2 (b) (c) FIG. 4: Relative weight of the P1 [Fig. 3(c)] as a function
P
of temperature. The thick continuous line (blue) is a guide
P1
) to the eyes. The size of the symbols represents maximum
s
nit0.0 uncertainty,estimatedbyGuassianfitsofdifferentspectraat
u 29 K each temperature.
b.
r 34 K
a
(
y 37 K withtwoGaussianfunctionsinthepre-edgepeakregion.
t P2
i
s The background is given by a constant arctangent func-
n
e 39 K
t tion approximating the atomic absorption. Two Gaus-
n
I 43 K sianfunctionsofconstantwidthandenergypositionwere
x5 enough to describe the background subtracted pre-edge
48 K
peakP.Wehaveshownthezoomedabsorptionspectrum
in the pre-edge peak region containing the background
7112 7116 7112 7116
function along with the two Gaussian components (P1
Energy (eV)
and P2) of the background subtracted spectrum and the
fit as Fig. 3(c).
FIG. 3: (a) Partial fluorescence yield (PFY) XAS spectra Let us discuss the possible origin of the two Gaussian
measuredattheFeK-edgeofK Fe Se atvarioustemper- peaks P1 and P2. It is known that Fe K-edge absorp-
x 2−y 2
atures. (b) A zoom over the pre-peak region with spectral tion pre-peak is sensitive to the Fe spin state in iron
differences across the transition temperature (multiplied by
complexes [33]. In addition, earlier high pressure XAS
five) with respect to a spectrum measured below the transi-
experiments on FeSe and Rb Fe Se have revealed a
tion. (c)Deconvolutedabsorptionpre-edgepeakintwocom- x 2−y 2
crossover from HS Fe2+ to LS Fe2+, evidenced by an ap-
ponents (P1 and P2) after subtracting a constant arctangent
background from the normalized absorption (the latter are pearanceofaLSFe2+ peakstructureat∼1eVhigherin
also shown). energythantheHSFe2+peak [34–36]. Wearealsoaware
of the fact that different phases coexist in the supercon-
ducting K Fe Se : (i) a majority semiconducting and
x 2−y 2
in the Fe 4p-Se 5d hybridization in K Fe Se [31, 32]. magnetic 245-phase with iron vacancy order, and; (ii) a
x 2−y 2
Ontheotherhand,pre-edgepeakexhibitssmalltemper- metallic and nonmagnetic disordered 122- phase. The
ature dependence, indicating change in the unoccupied coexisting phases are characterized by slightly different
Fe 3d-Se 4p electronic states [31, 32] as a function of Feoxidationstates, i.e., 2+, forthe245, andslightlyless
temperature. than 2+, for the 122 phase [21]. Small differences in the
In order to describe the temperature evolution of the oxidation state are hardly detectable by the main ris-
electronic states near the Fermi level, Fig. 3 (b) shows ing absorption edge that appears to be influenced by the
difference spectra across T ∼32 K with respect to a glassy local structure [30]. The 245 and the 122 phases
c
spectrum measured (at 23 K) well below T . The differ- havesimilarstructurewiththelatterbeingslightlycom-
c
ence spectra reveal spectral weight redistribution from pressed in the plane and expanded in the out of plane
the lower energy side to the higher energy side of the [8]. Thesephasesalsohavedifferentiron spinstate [24],
pre-edge peak. To evaluate the temperature dependent withthemajorityphasebeinginHSFe2+ statewhilethe
redistribution of the spectral weight, we have performed minority phase appears to be in LS state. Therefore, we
acurvefittingofthebackgroundsubtractedXASspectra canassignqualitativelythetwoGaussiancomponentsP1
5
andP2toHSFe2+ majorityphaseandLSFe2+ minority
(a) 300 K < T < 520 K (b) 180 K < T < 300 K
phase. Considering the fact that the minority and ma-
jority phases are known to contribute about 10-20% and
80-90% respectively, the spectral weight of the majority
phase (minority phase) with HS (LS) Fe2+ state count- LS
ingabout78%(22%)atRTisinfairagreementwiththe HS HS
present assignment.
Temperature evolution of the relative spectral weight
of the majority phase (defined as the ratio of the inten-
sity of P1 to the total intensity, i.e., P1+P2) is shown in (c) 32 K < T < 180 K (d) T < 32 K
Fig. 4. The HS phase tends to decrease with decreas-
ing temperature from 300 K to 180 K, revealing a cusp-
like anomaly around 180 K with maximum weight being SC
∼80% before a sharp drop to ∼70% around the T . On HS HS
c
the other hand, with increasing temperature from 300 K
the spectral weight of the HS phase gradually decreases
with a small discontinuity above the phase separation
temperature of ∼520 K and iron vacancy disorder tem-
perature of ∼580 K. In general, temperature evolution
of the relative weight of the majority phase appears con- FIG. 5: (a) Fe2+ HS domains gradually grow with cooling
sistent with the IAD representing variation of the local from520Kto300K.(b)FilamentaryFe2+ LSdomainises-
Fe moment µ [Fig. 1]. This justifies the assignment of tablishedwithcoolingfrom300Kto180K.(c)Theinterface
the P1 and P2 due to the two coexisting phases, charac- region between the Fe2+ HS and Fe2+ LS domains becomes
terized by HS and LS Fe2+ states, with HS phase con- magneticwithcoolingfrom180Kto32K.(d)Themagnetic
interfacestateturnsintothesuperconductingstatewithFe2+
tributing about 70-80% of the total while the remaining
LS.
phase should be due to LS Fe2+ phase. This affirma-
tion agrees with different experimental studies showing
that the magnetic phase is the majority phase while the Fe2+ LS fraction gradually increases, likely to be due to
minority phase should be one that becomes supercon- chemicalpressurefromthemajorityphase,andtheFe2+
ducting by cooling [4, 14, 15]. Again, it is interesting HS fraction decreases [Fig. 5(b)], which is also consis-
to note that, by cooling from 600 K to 400 K, the sys- tent with the decrease of the IAD. These behaviors are
tem does not recover the initial state. Indeed, after the basicallyconsistentwiththepreviousworksonelectrical
annealing, the discontinuity across Tc disappears, likely transportshowinganomaloustemperaturedependentre-
to be due to suppression of superconductivity or highly sistivity[1–3]. Theresultsarealsoinagreementwiththe
reduced volume fraction of the superconducting phase. magnetic [4, 7, 16, 17], optical [18, 19] and microscopy
Let us attempt to understand the XES and XAS find- measurements [22]. The most striking result is the in-
ings shown in Fig. 2 and Fig. 4, revealing an abnor- crease of the Fe2+ HS fraction with cooling from 180 K
mal evolution of local Fe magnetic moment and different althoughthesystemexhibitsagoodmetallicbehaviorin
electronic phases. We start our discussion for the cool- the temperature range. This observation suggests that
ing cycle from 300 K. It is clear that the local moment the Fe2+ LS state at the interface region again tends to
shows a decrease [Fig. 2], consistent with the decrease be magnetic but stay metallic as schematically shown in
of the HS Fe2+ phase [Fig. 4]. With further cooling, the Fig. 5(c). One possible explanation is that the magnetic
Fe magnetic moment shows an increase before showing fluctuation and local spin moment are induced by or-
a sharp drop at the superconducting transition. This is bitalselectiveMotttransitioninthistemperaturerange.
againconsistentwiththeevolutionoftheHSFe2+ phase SuchaphasehasbeenproposedrecentlyforK Fe Se
x 2−y 2
fraction. However, the question is what is the origin of on the basis of angle-resolved photoemission results [37].
such an anomalous behavior of the phase fractions. Thus, the apparent increase of the local magnetic mo-
Below ∼580 K, the system undergoes ordering of the ment below 180 K can be explained by the orbital se-
Fe vacancy and nucleation of the HS Fe2+ phase with lective localization of the Fe 3d electrons and resulting
phase separation at ∼520 K. In going from 520 K to strong magnetic fluctuations. This assumption is con-
300 K, the antiferromagnetic order is established and, sistent with the recent high pressure measurements un-
consequently, the IAD and Fe2+ HS fraction gradually derlining the importance of the insulating texture in the
increase with cooling [Fig. 5(a)]. Around 300 K, the superconductivity of A Fe Se [38]. Across T , the in-
x 2−y 2 c
filamentary and metallic LS Fe2+ state starts to grow terface metallic region with the localized magnetic mo-
at the boundary between the neighboring Fe2+ HS do- ment turns into the superconducting LS state, connect-
mains. In going from 300 K to 180 K, the filamentary ing all filamentary metallic phase, as illustrated in Fig.
6
5(d). It is worth mentioning that the Fig. 5 provides [6] W. Bao, Q.Z. Huang, G.F. Chen, M. A. Green, D.M.
a likely interpretaion of the anomalous behavior of local Wang, J.B. He, and Y.-M. Qiu, Chin. Phys. Lett. 28,
magnetic moment [Fig. 2] and pre-peak intensity [Fig. 086104 (2011).
[7] F. Ye, S. Chi, Wei Bao, X. F. Wang, J. J. Ying, X. H.
4] as a function of temperature. However, it is difficult
Chen,H.D.Wang,C.H.Dong,andMinghuFang,Phys.
to argue that P1 and P2 in Fig. 3 to provide quantita-
Rev. Lett. 107, 137003 (2011).
tive estimation of the HS and LS phases since the iron
[8] A.Ricci,N.Poccia,G.Campi,B.Joseph,G.Arrighetti,
vacancyconfigurationinthemajorityphaseitselfevolves L. Barba, M. Reynolds, M. Burghammer, H. Takeya, Y.
with temperature. Mizuguchi, Y. Takano, M. Colapietro, N. L. Saini, and
A. Bianconi, Phys. Rev. B 84, 060511(R) (2011).
[9] Z. Wang, Y. J. Song, H. L. Shi, Z. W. Wang, Z. Chen,
H. F. Tian, G. F. Chen, J. G. Guo, H. X. Yang, and J.
CONCLUSIONS
Q. Li, Phys. Rev B 83, 140505(R) (2011).
[10] F. Chen, M. Xu, Q. Q. Ge, Y. Zhang, Z. R. Ye, L. X.
In summary, we have employed high energy XES and Yang,JuanJiang,B.P.Xie,R.C.Che,M.Zhang,A.F.
XAS to investigate local magnetic moment and coexist- Wang,X.H.Chen,D.W.Shen,X.M.Xie,M.H.Jiang,
ingelectronicphasesinK Fe Se asafunctionoftem- J. P. Hu, D. L. Feng, Phys. Rev. X 1, 021020 (2011).
x 2−y 2
perature, across the superconducting and phase separa- [11] J.Guo,X.Chen,C.Zhang,J.Guo,X.Chen,Q.Wu,D.
tion temperatures. We have found coexisting Fe2+ elec- Gu, P. Gao, X. Dai, L. Yang, H. Mao, L. Sun, Z. Zhao,
Phys. Rev. Lett. 108, 197001 (2012).
tronic phases with the glassy character of K Fe Se
x 2−y 2 [12] D.M.Wang,J.B.He,T.-L.Xia,andG.F.Chen,Phys.
compound. The results provide a clear evidence of the
Rev. B 83, 132502 (2011).
interplay between the electronic andmagnetic degrees of [13] A. M. Zhang, J. H. Xiao, J. B. He, D. M. Wang, G. F.
freedom for the superconductivity. Across Tc, the LS Chen,andQ.M.Zhang,Phys.Rev.B85,024518(2012).
phaseincreasesattheexpensesoftheinterfacephasebe- [14] Y. J. Yan, M. Zhang, A. F. Wang, J. J. Ying, Z. Y. Li,
tweenLSandHSphases. TheHStoLSfractionexchange W.Qin, X.G.Luo, J.Q.Li, JiangpingHu, X.H.Chen,
Sci. Rep. 2, 212 (2012).
acrossT issuppressedafterannealingdespiteasubstan-
c
[15] F. Han, B. Shen, Z.-Y. Wang, H.-H. Wen, Philos. Mag.
tial presence of the LS configuration, likely to be due to
92, 2553 (2012).
absence of a well defined interface phase. The results
[16] Y. Texier, J. Deisenhofer, V. Tsurkan, A. Loidl, D. S.
suggest that a delicate interplay of local strain between Inosov,G.Friemel,andJ.Bobroff,Phys.Rev.Lett.108,
the Mott insulator and metallic filaments with an inter- 237002 (2012).
facephaseisthekeyforthesuperconductingphenomena [17] Z. Shermadini, H. Luetkens, R. Khasanov, A. Krzton-
of inhomogeneous K Fe Se system. Maziopa,K.Conder,E.Pomjakushina,H-H.Klauss,and
x 2−y 2
A. Amato, Phys. Rev. B 85, 100501 (2012).
[18] C. C. Homes, Z. J. Xu, J. S. Wen, and G. D. Gu, Phys.
Rev. B 86, 144530 (2012).
ACKNOWLEDGMENTS
[19] R.H.Yuan,T.Dong,Y.J.Song,P.Zheng,G.F.Chen,
J.P.Hu,J.Q.LiandN.L.Wang,SciRep.2,221(2012).
We thank A. Sulpice for the magnetic susceptibility [20] V. Ksenofontov, G. Wortmann, S. A. Medvedev, V.
measurements on the annealed sample at the end of the Tsurkan, J. Deisenhofer, A. Loidl, and C. Felser, Phys.
thermal cycles. One of us (T.M.) would like to acknowl- Rev. B 84, 180508 (2011).
[21] D. P. Shoemaker, D. Y. Chung, H. Claus, M. C. Fran-
edge hopsitality at the Sapienza University of Rome.
cisco, S. Avci, A. Llobet, and M. G. Kanatzidis, Phys.
This work is partially supported by PRIN2012 (grant
Rev. B 86, 184511 (2012).
numbre 2012X3YFZ2) of MIUR, Italy.
[22] M. Bendele, A. Barinov, B. Joseph, D. Innocenti, A.
Iadecola, H. Takeya, Y. Mizuguchi, T. Takano, T. Noji,
T. Hatakeda, Y. Koike, M. Horio, A. Fujimori, D. Oot-
suki, T. Mizokawa, and N. L. Saini, Sci. Rep. 4, 5592
(2014).
[1] J. Guo, S. Jin, G.Wang, S.Wang, K. Zhu, T. Zhou, M. [23] M.Oiwake,D.Ootsuki,T.Noji,T.Hatakeda,Y.Koike,
He, and X. Chen, Phys. Rev. B 82, 180520 (2010). M. Horio, A. Fujimori, N. L. Saini, and T. Mizokawa,
[2] Y. Mizuguchi, H. Takeya, Y. Kawasaki, T. Ozaki, S. Phys. Rev. B 88, 224517 (2013).
Tsuda,T.Yamaguchi,andY.Takano,,Appl.Phys.Lett. [24] L.Simonelli,N.L.Saini,M.MorettiSala,Y.Mizuguchi,
98, 042511 (2011). Y. Takano, H. Takeya, T. Mizokawa, and G. Monaco,
[3] J. J. Ying, X. F. Wang, X. G. Luo, A. F. Wang, M. Phys. Rev. B 85, 224510 (2012).
Zhang, Y. J. Yan, Z. J. Xiang, R. H. Liu, P. Cheng, G. [25] C. Ponchut, J. M. Rigal, J. Cl´ement, E. Papillon, A.
J. Ye, and X. H. Chen, Phys. Rev. B 83, 212502 (2011). Homs, and S. Petitdemange, Journal of Instrumentation
[4] M.H. Fang, H.D. Wang, C.H. Dong, Z.J. Li, C.M. Feng, 6, C01069 (2011).
J. Chen, and H. Q. Yuan, Europhys. Lett. 94, 27009 [26] P. Glatzel and Uwe Bergmann, Coordination Chemistry
(2011). Reviews 249, 65 (2005).
[5] D.H. Ryan, W. N. Rowan-Weetaluktuk, J.M. Cadogan, [27] R.D.Cowan,TheTheoryofAtomicStructureandSpec-
R.Hu,W.E.Straszheim,S.L.Bud’ko,andP.C.Canfield, tra, University of California Press, Berkeley, (1981).
Phys. Rev. B 83, 104526 (2011). [28] G.Vanko´,T.Neisius,G.Molna´r,F.Renz,S.Ka´rpa´ti,A.
7
Shukla, and F. M. F. de Groot, J. Phys. Chem. B 110, Ishii,K.D.Tsuei,E.Huang,andT.J.Yang,Phys.Rev.
11647 (2006). B 84, 125117 (2011).
[29] H. Gretarsson, S. R. Saha, T. Drye, J. Paglione, Jungho [35] R.S.Kumar,Y.Zhang,Y.Xiao,J.Baker,A.Cornelius,
Kim,D.Casa,T.Gog,W.Wu,S.R.Julian,andYoung- S. Veeramalai, P. Chow, C. Chen, and Y. Zhao, Appl.
June Kim, Phys. Rev. Lett. 110, 047003 (2013). Phys. Lett. 99, 061913 (2011).
[30] A. Iadecola, B. Joseph, L. Simonelli, A. Puri, Y. [36] M.Bendele,C.Marini,B.Joseph,G.M.Pierantozzi,A.
Mizuguchi, H. Takeya, Y. Takano, N.L. Saini, J. Phys.: S. Caporale, A. Bianconi, E. Pomjakushina, K. Conder,
Condens. Matter 24, 115701 (2012). A.Krzton-Maziopa,T.Irifune,T.Shinmei,S.Pascarelli,
[31] L. Simonelli, N. L. Saini, Y. Mizuguchi, Y. Takano, T. P.Dore,N.L.Saini,andP.Postorino,Phys.Rev.B88,
Mizokawa, G. Baldi and G. Monaco, J. Phys. Condens. 180506 (2013).
Matter 24, 415501 (2012). [37] M.Yi,D.H.Lu,R.Yu,S.C.Riggs,J.-H.Chu,B.Lv,Z.
[32] B. Joseph, A. Iadecola, L. Simonelli, Y. Mizuguchi, Y. K. Liu, M. Lu, Y.-T. Cui, M. Hashimoto, S.-K. Mo, Z.
Takano,T.Mizokawa,andN.L.Saini,J.Phys.Condens. Hussain, C.W. Chu, I. R. Fisher, Q. Si, and Z.-X. Shen,
Matter 22, 485702 (2010). Phys. Rev. Lett. 110, 067003 (2013).
[33] T. E. Westre, P. Kennepohl, J. G. DeWitt, B. Hedman, [38] P. Gao, R. Yu, L. Sun, H. Wang, Z. Wang, Q. Wu, M.
K. O. Hodgson, and E. I. Solomon, J. Am. Chem. Soc. Fang,G.Chen,J.Guo,C.Zhang,D.Gu,H.Tian,J.Li,
119, 6297 (1997). J. Liu, Y. Li, X. Li, S. Jiang, K. Yang, A. Li, Q. Si, and
[34] J. M. Chen, S. C. Haw, J. M. Lee, T. L. Chou, S. A. Z.X. Zhao, Phys. Rev. B 89, 094514 (2014).
Chen, K. T. Lu, Y. C. Liang, Y. C. Lee, N. Hiraoka, H.