Table Of Content57Fe Mössbauer study of Lu Fe Si iron silicide superconductor
2 3 5
Xiaoming Ma1,2, Sheng Ran2, Hua Pang1, Fashen Li1 Paul C. Canfield2 and Sergey L. Bud’ko2∗
1Institute of Applied Magnetics, Key Laboratory for Magnetism and Magnetic Materials of the Ministry of Education,
Lanzhou University, Lanzhou, Gansu Province, 730000, China.
2Ames Laboratory, US DOE, and Department of Physics and Astronomy, Iowa State University, Ames, Iowa 50011, USA
(Dated: January 30, 2015)
With theadventof Fe-Asbased superconductivityit hasbecome important to studyhow super-
conductivitymanifestsitselfindetailsof57FeMössbauerspectroscopyofconventional,Fe-bearing
5 superconductors. To this end, the iron-based superconductor Lu2Fe3Si5 has been studied by 57Fe
1
Mössbauer spectroscopy over the temperature range from 4.4 K to room temperature with par-
0
2 ticular attention to the region close to the superconducting transition temperature (Tc = 6.1 K).
Consistent with thetwo crystallographic sites for Fe in this structure, the observed spectra appear
n tohaveapatternconsistingoftwodoubletsoverthewholetemperaturerange. ThevalueofDebye
a temperaturewas estimated from temperaturedependenceof theisomer shift andthetotal spectral
J areaandcomparedwiththespecificheatcapacitydata. Neitherabnormalbehaviorofthehyperfine
9 parameters at or near Tc, nor phonon softening were observed.
2 PACS numbers: 74.70.Dd, 76.80.+y
] Keywords: A.superconductors; C.Mössbauerspectroscopy; D.specificheat
i
c
s
- I. INTRODUCTION spectroscopytopurelymagneticorpurelysuperconduct-
l
r ing origins [12, 14–16].
t
m Since the discovery of Mössbauer effect spectroscopy In order to study possible variations of hyperfine pa-
. more than half a century ago [1], superconductivity has rameterscausedbythetransitionfromthenormaltothe
at been one of the states that has been investigated by superconducting state in a conventionalsuperconductor,
m this technique [2]. Although Mössbauer spectroscopy is we revisitedlutetium-iron-silicide, Lu2Fe3Si5. Lu2Fe3Si5
widelyacceptedasoneofthemostsensitivetechniquesin is a stoichiometric, Fe-containing, superconductor with
-
d terms of energy resolution, it has not contributed signif- relatively high Tc ≈ 6 K [17]. In a previous Mössbauer
n icant insight to studies of conventional superconductors. study [18], the non-magnetic nature of Fe was already
o After the discovery of cuprate high temperature super- confirmed. Hence, Lu2Fe3Si5 can be considered as an
c
conductors(HTSC), Mössbauerspectroscopywaswidely ideal compound to investigate the variation of hyperfine
[
usedforstudiesofthesematerialsandreportsofobserva- parameters caused only by the superconducting transi-
2 tion of some anomalies in the spectral parameters in the tionwithoutanycomplicationsassociatedbytheabsence
v
vicinityofthesuperconductingcriticaltemperature(T ) ofMössbauernucleusand/orthe presenceofmagnetism.
4 c
[3–7] were published. Due to the absence of commonly To the best of our knowledge no detailed, temperature
5
2 availableMössbauernuclides in the cuprates,moststud- dependent, 57Fe Mössbauer spectroscopy measurements
7 ies were accomplished either by partial substitution of wereperformedonthismaterialsofar,andourgoalisto
0 copperatomsby57Feand/or119Sn,orbyusingresonant shed some light on the applicability of Mössbauer spec-
. isotopes of the rare earth metals, like 151Eu, which in- troscopy for studies of conventional, albeit multigap su-
1
0 creases the degree of difficulty of the measurements and perconductors [19].
5 reduces the clarity of the results [3, 5, 8, 9]. Recently,
1 the discovery of iron-based superconductors, that natu-
v: rally contain the common Mössbauer nuclide, 57Fe, has II. EXPERIMENTAL DETAILS
i triggered intense Mössbauer studies of these supercon-
X
ductors [10–15]. Superconductivity in iron-pnictides is
r usuallyachievedbydopingamagneticparentcompound PolycrystallinesamplesofLu2Fe3Si5 werepreparedby
a arc melting constituent elements with the nominal com-
with electrons or holes, or by application of chemical or
physicalpressure,andthereby,suppressingthe magnetic position of Lu2Fe3.32Si5.26 (corresponding to Lu2Fe3Si5
order,suggesting that superconductivity and magnetism + Fe0.32Si0.26) inZr-getteredAr atmosphere. Extra iron
are closely related in this system. Although there are was added to suppress the formation of a Lu2FeSi4 sec-
ond phase and extra silicon was added to compensate
some studies oniron-basedsuperconductors,whichstate
apparent loss during the arc melting. To ensure the ho-
thatMössbauerspectralparametersshowanomaliesnear
mogeneity of the sample, the arc melting was repeated
T , in these materials it is hard to attribute the varia-
c
iterativelyafterflippingofthemeltedandresolidifiedin-
tion of the hyperfine parameters observed by Mössbauer
got, for more than ten times. The weight loss was about
0.26%. Thearc-meltedingotwasthensealedinanamor-
phous silica tube, under a partial pressure of argon, and
∗ Correspondingauthor: [email protected] annealed at 1050 ℃ for 12 days.
2
Powder X-ray diffraction (XRD) was performed us-
ing a Rigaku Miniflex diffractometer with Cu Kα radia- 10000
obs
tionatroomtemperature(RT).The powderX-rayspec- calc
tra of the samples were refined by Rietveld analysis us- bkg
diff
ing the EXPGUI software [20]. Magnetic measurements
1000 Lu2Fe3Si5
wereperformedusingaQuantumDesignMagneticProp- s
nt alpha-Fe
erty Measurement System SQUID magnetometer, spe- u
o
cific heat capacity was measured in a Quantum Design C
Physical Property Measurement System. 100
Mössbauer spectroscopy measurements were per-
formed using a SEE Co. conventional constant accelera-
tiontypespectrometerintransmissiongeometrywithan
10
57Co(Rh) source, which had an initial intensity 50 mCi, 0
keptatRT.Theabsorberwaspreparedinapowderform
(10 mg of natural Fe/cm2) by grinding of approximately -500
25 30 35 40 45 50 5 5 60 65 70 75 80
175mgpieceofthearc-meltedandannealedbutton. The
2 (deg.)
absorberholdercomprisedtwonestedwhite Delrincups.
FIG. 1. (Color online) Rietveld refinement of the powder
The powder was placed uniformly on the bottom of the
XRD spectrum of Lu2Fe3Si5. Measured (black cross), calcu-
larger cup and was held in place by a smaller cup. The
lated intensities (red line) and background (green line) and
absorber holder was locked in a thermal contact with a
difference curve (blue line) are shown. Vertical bars at the
copper block with a temperature sensor and a heater,
bottom indicate the positions of the Bragg reflections. The
and aligned with the γ - source and detector. The ab- peaks of unknownphase are marked by theblue arrows.
sorber was cooled to a desired temperature using Janis
model SHI-850-5 closed cycle refrigerator (with vibra-
tions damping) that has long-termtemperature stability
better than 0.1 K at low temperature. The driver veloc-
1.005
ity was calibrated by α-Fe foil and all isomer shifts (IS)
are quoted relative to the α-Fe foil at RT. At first, spec-
tra with maximum velocity 6 mm/s and 3 mm/s were a.u.)
both measured at RT to check that no iron-containing n (
o1.000
impurity can be seen in the Mössbauer spectra. Then, pti
or
three rounds of measurements, progressively focusing in bs
A
on temperatures near Tc =6.1 K, were carried out: col- e
v
lecting 24 h with maximum velocity 2 mm/s from 4.3 ati0.995 1%
K to 293.8 K (S1); collecting 48 h with maximum ve- Rel
locity 3 mm/s from 4.4 K to 10 K (S2); collecting 48
h with maximum velocity 3 mm/s from 4.7 K to 6.4 K
-6 -4 -2 0 2 4 6
(S3). All the Mössbauer spectra were fitted by the com- 0.990
-6 -4 -2 0 2 4 6
mercial software package MossWinn [21], in which the
Velocity (mm/s)
standard error of parameters can be estimated either by
calculating and inverting the curvature matrix of the χ2
FIG.2. (Coloronline)Expandedviewofthebackgroundpart
withrespecttothefitparameters,orbytheMonteCarlo
of the RT Lu2Fe3Si5 Mössbauer spectrum in large velocity
method. The standarderrorof IS, quadroupole splitting scale. Red arrows show the expected peaks positions for α-
(QS) andline width (Γ) inthis workwere obtainedfrom Fe.The inset is thefull view of thespectrum.
the curvature matrix, while the error the area under the
spectra here were obtained by Monte Carlo method by
iterating 100 times.
known phase. A 1.8 wt.% of α-Fe corresponds to 6.7
% of Fe atoms in the α-Fe form in the ground sample,
which,forun-enrichedFe, is belowthe resolutionof57Fe
III. RESULTS AND DISCUSSION Mössbauer spectroscopy to detect α-Fe. As can be seen
in Fig. 2, there are no peaks associated with the α-Fe
A. Structure and superconductivity or with the other impurity in the Mössbauer spectrum,
which means the unknown phase is either iron-free or, if
The XRD pattern of the Lu2Fe3Si5 polycrystalline containsiron,isbelowtheresolutionlimit. Consequently,
sample is presentedin Fig. 1. The majority ofthe peaks theMössbauerspectracanbeanalyzedasasingle-phase
matchtothetetragonalstructurewiththeP4/mncspace (Lu2Fe3Si5) spectra.
group. The Rietveld refinement results in an estimate of The superconductivity of the sample is confirmed by
∼ 1.8(1) wt.% α-Fe impurity and a trace amount of un- the dc susceptibility measurement in a magnetic field of
3
10 Oe. As shown in Fig. 3 (a), the susceptibility data
shows diamagnetic signal below ∼6.2 K. The transition
issharpwithawidthoflessthan0.4K.Fig. 3(b)shows
0 Lu2Fe3Si5 the temperature dependent specific heat capacity (Cp).
) Asuddenjumpcausedbysuperconductingtransitioncan
e
O be observed below 6.3 K on cooling, the transition tem-
ol perature is close to that obtained from the susceptibility
m
u/ data. Basedonthese two thermodynamicmeasurements
m -4 we take T = 6.1±0.1 K.
H (e Cp/T iscalso plotted as a function of T2 in Fig. 4. A
M/ linearfitabovethe superconductingtransitionyieldsthe
values of γ (γ T is the electronic contribution to spe-
n n
-8 cific heat capacity) and β (β T3 represents the phonon
n n
(a) contribution)ofγ = 24.6(2)mJ/molK2, β = 0.287(2)
n n
4
mJ/mol K , which are very close to the previously re-
500
portedvalues[22]. Fromtheβ value,wecanestimatethe
value of the Debye temperature (Θ ) using the relation:
D
400 Θ = (12π4Nrk /5β)1/3, where N is Avogadro’s num-
D B
K) ber, r is the number of atoms per formula unit, and kB
ol 300 is the Boltzmann’s constant. We further obtained ΘD is
m
408K.Thevalueofthenormalizedspecific-heatjumpat
J/
m T , ∆C/γ T , is ≈1.06,a value that is smaller than the
c n c
(p200 BCSvalue of1.43,but consistentwith the previouslyre-
C
portedvalue1.05[22]. Fromtheabovecharacterizations,
we can conclude that our sample is a bulk superconduc-
100
tor and can be considered as single phase for Mössbauer
(b)
measurements and analysis.
0
2 4 6 8 10
Temperature (K) B. Mössbauer results and discussion
FIG. 3. (Color online) (a) Temperature dependent of zero 1. Symmetry of Fe sites and choice of the model
fieldcooledmagneticsusceptibilityofLu2Fe3Si5(H=10Oe);
(b)thetemperaturedependentlowtemperaturespecificheat
capacity. Vertical dashed line marksTc =6.1 K. In the Lu2Fe3Si5 crystal structure, there are two,
nonequivalent, Fe positions, FeI and FeII of 1 : 2 oc-
cupation. FeI atoms are located at the 4d sites, which
form 1D chains along the c axis. FeII atoms are located
at 8h sites, which form squares with planes perpendic-
ular to the c axis. In each Fe position, the Fe atom is
locatedinapolyhedronformedbySiatoms. FeI hasfour
60 Lu2Fe3Si5
Si atoms at a distance of 2.31 Å which form an irregular
tetrahedronandtwoSiatomsatadistanceof2.54Åwith
2 K) the nearestFe-Fe distanceis 2.67Å.The FeII hasfourSi
ol 40 atomsata distanceof2.34Å whichareinthe sameface
m
J/ and form a quadrangle. On each side of the face, there
m
T( are two Si atoms with the 2.35 Å Fe-Si distances. The
C/p20 nearest Fe-Fe distance for FeII is 2.71 Å.
The anisotropic environments of the Fe atoms ensure
nonzeroelectricfieldgradient(EFG)tensoratbothsites.
Hence,in analyzingthe data,twodoublets areexpected.
00 20 40 60 80 100 All the 57Fe Mössbauer spectra of Lu2Fe3Si5 over the
2 2
T (K) whole observedtemperature range share similar spectral
shapeswithaclearquadrupolesplittingasshowninFig.
5 (a). Averysmallasymmetryandsmallshoulderswere
FIG. 4. (Color online) The specific heat capacity divided by
temperature, Cp/T, as a function of T2 for Lu2Fe3Si5. The observed,which suggest that at least two subspectra are
dashedlinerepresentsthelinearfittothedatainthenormal neededto resolvethe spectra. However,due to the small
state as described in thetext. splittingandconsequentlypoorresolutionofthespectra,
the data can be analyzed with more than one set of pa-
4
TABLEI.Hyperfineparametersobtainedbyfittingusingdif-
ferent models as discussed in the text. IS is the isomer shift,
QSisthequadrupolesplittingandIistherelativeareaofthe
two subspectra.
T model site IS QS I χ2
(a) RT
K mm/s mm/s %
RT 1 4d 0.129(3) 0.311(1) 39.4 1.63
8h 0.278(3) 0.296(1) 60.6 Lu2Fe3Si5
RT 2 4d 0.2157(9) 0.513(6) 36.4 0.98
8h 0.2206(6) 0.248(8) 63.6 u.)
4.4 2 4d 0.320(2) 0.523(4) 52.8 1.29 a.
8h 0.330(1) 0.202(3) 47.2 n ( (b) RT Model 1
o
pti
r
o
s
rameters. InthepreviousMössbauerstudiesofR2Fe3Si5 b
A
(R = rareearth), both (i) one doublet, and(ii) two dou- e
blets with fixed arearatio were employedto fit the spec- v
tra [18, 23]. In our approachto the fitting, two doublets ati (c) RT Model 2
el
without any restriction are used. In model 1, the two R
subspectrahaveclosevaluesofQSbutobviouslydifferent
valuesofIS,whichissimilartothereportedSc2Fe3Si5 fit
result [24]. In model 2, the two subspectra have similar
values of IS, but distinct QS values. Two sets of param-
eters yielding fits of acceptable quality can be obtained,
(d)
thecorrespondingparametersoftheRTspectrumfitsare
listed in the Table 1. The RT spectrum fitted using our
two models is shown in Fig. 5 (b) and (c). The relative 4.4 K Model 2
2%
area of two subspectra in model 2 are closer to the the-
oretical value 1 : 2 and χ2 values are closer to 1, so we
have chosen the model 2 to fit all the collected spectra.
-2 0 2
As an example, the fit using model 2 of the spectrum
measuredat4.4 Kfromthe S2 setis showninFig. 5 (d)
Velocity (mm/s)
and the corresponding parameters are listed in Table 1.
FIG. 5. (Color online) 57Fe Mössbauer spectra of Lu2Fe3Si5.
(a)the spectrum at RT; the results of the fits using model 1
2. Hyperfine parameters andmodel2areshownin(b)and(c),respectively;(d)isthe
spectra collected at 4.4 K and fitted using the model 2. The
model 1 and model 2are described in detail in thetext.
Fig. 6 summarizes the variation of hyperfine param-
eters of the spectra with temperature. The plots in the
right column present an expanded view of the low tem-
chemical shift should not depend on temperature. The
perature range.
main contribution to this variation is from the second-
The IS of Lu2Fe3Si5 plotted as a function of tempera- orderDoppler shift, which is usually described by Debye
ture is shown in Fig. 6 (a) and (b). The IS values of the model:
4dand8hsitesat294Kare0.225(2)mm/sand0.227(1)
mm/s, respectively, which are slightly larger than the 9k T T ΘD/T x3dx
B 3
IS(T)=IS(0)− ( ) , (1)
typicalvaluesforiron-siliconcompoundsinwhichFecar- 2 Mc ΘD Z0 ex−1
ries no moment[25, 26], but areabout 0.2mm/s smaller
than that in the iron-pnictide compounds [10–12]. The wherec is the velocity oflight,M is the mass ofthe 57Fe
temperaturedependenciesoftheIScorrespondingtothe nucleus, and IS(0) is is the temperature-independent
two sites are very similar, and no anomalies can be ob- part, i.e. the chemical shift. A fit with Eq. (1) to the
served around T (Fig. 6(b)). The IS values obtained data of S1 shown in Fig. 6 (a) yields Θ = 517(18) K
c D
from the fits includes contributions from both the chem- and 545(16) K for 4d and 8h sites, respectively.
ical shift and the second-order Doppler shift, which is The quadrupole splittings in Lu2Fe3Si5 at the 4d and
knowntoincreaseconvexlyupondecreasingtemperature, 8h sites are 0.50(1) mm/s and 0.19(1) mm/s at 294 K,
duetogradualdepopulationoftheexcitedphononstates. respectively. The magnitude of the QS is proportional
However, it should be constant at low temperature, be- to the z component V of the EFG tensor, which is
zz
causeofthequantummechanicalzero-pointmotion. The composed of two contributions: (V ) , from the lig-
zz lig
5
0.35 (a) (b) 0.36
)
s
m/ 0.34
0.30
m
( S1 4d S1 8h
S
I S2 4d S2 8h 0.32
0.25
S3 4d S3 8h
0.30
0.55 (c) (d) 0.55
)
s
m/ 0.50 0.50
m 0.45 0.45
(
S
0.25 0.25
Q
0.20 0.20
0.15 0.15
) 0.35 (e) (f) 0.35
s
m/
m 0.30 0.30
(
0.25 0.25
0.20 0.20
)
%
70 70
( (g) (h)
a
e 60 60
r
A
e 50 50
v
ati 40 40
el
R
30 30
10 100 4 5 6 7
Temperature (K) Temperature (K)
FIG. 6. (Color online) Temperature dependence of parameters derived from fitting the of Lu2Fe3Si5 Mössbauer spectra with
model2asdescribedinthetext. (a)Isomershift,(c)quadrupolesplitting,(e)linewidthand(g)relativeareaoftwosubspectra.
S1,S2, S3- mark threemeasurement sets, theparameters are givenfor twoFesites, 4d and8h (seethetextfor moredetails).
The lines in panel (a) are fits to the Eq. (1). The figures in the right column are the expansion one of the low temperature
region of the corresponding left figure. The temperature stability for each of the measurements in the right column is better
than that presented bythesize of the symbols.
57
andchargesaroundtheMössbauernucleus,and(V ) , charge distribution around the Fe sites.
zz val
fromthevalenceelectronsofMössbauernucleus. Usually,
(V ) is small and weakly dependent on temperature, In order to get a better understanding of the elec-
zz lig
whereas (V ) is strongly temperature dependent. As tronic origin of EFG at the Fe sites in Lu2Fe3Si5, a
zz val
canbe seenfromFig. 5(c). the QSoftwositesareboth first-principles calculation was performed using the full-
potential linearized augmented plane wave method as
almosttemperatureindependent,whichindicatestheQS
ismainlydeterminedbythecontributionfromtheligand embodied in the WIEN2K [27, 28]. The generalizedgra-
dientapproximation(GGA)suggestedbyPerdew,Burke,
6
at the 4d and 8h sites should be 33.3% and 66.7%. At
RT the relative areas are close to expected value. How-
0.076
ever, as shown in Fig. 6 (e) as temperature decreases,
0.074 (a) the relative area values deviate from theoretical values
gradually, with the relative area for the 4d site becom-
s)0.072
m/ ing even larger than that for 8h site at low temperature.
m0.070
ne (0.068 D=385(8) K Tlinheiswpidhtehn,oΓm,efnoornthmeaydobueblreetlsatceodrrteospthoendvianrgiattoionthoefttwhoe
eli
s siteswithreducingtemperature. AsshowninFig. 6(g),
a0.066
B
a/ S1 Experimental Data the Γ of 4d site increases slightly at lower temperature,
e0.064 Fit
Ar but the Γ of 8h site is almost constant during the whole
0.062 temperature range. The slightly increase of Γ of 4d site
suggests the existence of instability at this lattice site at
0.060
low temperature.
0.058
10 100
1.05
3. The spectral area
(b)
1.00
S1
ea S2 In Mössbauer studies of superconductivity, the varia-
d Ar0.95 S3 tion of the total spectral area has also been the focus
e of discussion. There are number of reports showing a
z
mali0.90 1.00 decrease of the spectral area near Tc due to the soften-
or ing of lattice with the opening of superconducting gap
N
0.96 in cuprates [3–5] but to the best of our knowledge only
0.85 one report on iron-pnictide superconductors [32]. There
0.92 aretwokindsofbehaviorreported: arapiddecreasenear
0.80 T [3, 6, 7]; and a pit-like decrease either around T or
4 5 6 7 8 9 10 c c
at higher temperature, serving as a precursor to T [4].
c
0 50 100 150 200 250 300
Atthe sametime, somepublicationsnotedapoorrepro-
Temperature (K)
ducibilityofthoseobservationsinthecuprates[8,33]. In
our measurements, as can be seen in the Fig. 7, for the
FIG. 7. (Color online) (a) Temperature dependence of the S1, we did not observe any abnormal variation around
spectral area for S1 set of measurements on a semi-log scale. 6.1 K. To make sure we didn’t miss any minor variation
The solid line is a fit to Eq.(2), as explained in the text. (b) around T , we remeasured the Mössbauer spectra of the
Normalized to the values at ≈ 6 K, temperature dependent c
same sample between 4.4 K to 10 K with higher density
spectral area data for all three sets of measurements. Inset:
of the data points around 6.1 K for 48 h. Surprisingly,
enlarged low temperature part.
fordatasetS2,thereisasharp,6%,decreasearound5.5
K. Nevertheless, when we repeated the same measure-
ment again, data set S3, this phenomenon disappeared.
and Ernzerhof (PBE GGA) [29] was employed for the The sharp spectral area change seen in data set S2 is, in
exchange-correlationeffects. Once the electron densities our opinion, most likely an artifact. The feature in data
are calculated self-consistently and with high accuracy, set S2 occurs resolvably below T = 6.1 K with the very
c
the EFG tensor can be obtained from an integral over sharp and rather large jump occuring between spectra
the non-spherical charge density. The principal compo- taken at 5.3 K and 5.7 K (with spectra at 5.7 K, 6.0 K,
nent Vzz for the 4d site is 2.25×1021V/m2 and asym- and 6.3 K all having normalized areas near 1.0). If this
mforettrhyep8ahrasmiteetiesr-η1.=16×(V1x0x21−VV/ymy2)/aVnzdz =η =0.405.370;7t,hwehViczhz feeitahtuerrebwetewreeeanss5o.c7iaKtedanwdit6h.0TKc, iotrshboetuwldeehnav6e.0ocKcuarnedd
qualitativelyagreewith experimentalresults. It is found 6.3 K. In addition, there are no corresponding anoma-
that the p-p and d-d interactions mainly contribute to lies in data set S2’s isomer shift, quadrupole splitting
the EFG of Lu2Fe3Si5 and p- electrons is play a signif- or relative areas (as shown in Fig. 6). A simple expla-
icant role for states far from the Fermi energy whereas nation for such behavior could be, among others, some
the d-d interaction dominates around the Fermi energy, mechanical shift/rearrangement of the powder compos-
which is similar to what is found for the iron-pnictides ing the absorber. This observation gives a warning that
superconductors [30, 31]. even in the measurement of iron-containing stoichiomet-
Therelativeareasofthesubspectraaredeterminedby ric material, irreproducibilities/artifacts might exist and
the proportion of the Fe atoms on different lattice sites. should be addressed appropriately.
As mentioned above, for Lu2Fe3Si5, the theoretical rela- Finally, we also fitted the temperature dependence of
tive area of the two subspectra representing iron atoms areaunderthetwodoubletslineofS1measurementswith
7
Debye model: principlescalculationsyieldthevaluesofEFGthatqual-
itatively agree with the experiment. The Debye temper-
−3Eγ2 1 T 2 ΘD/T xdx ature was estimated by the temperature dependence of
f =exp[ { +( ) }], (2)
kBΘDMc2 4 ΘD Z0 ex−1 specific heat capacity, spectral area and IS. The ΘD ob-
tainedfromtemperaturedependenceofspectralareaand
where f is the recoilless fraction, which is proportional heat capacity are very similar, but about 140 K smaller
to the area for thin sample and Eγ is the γ-ray energy. than the value estimated by the IS variation with tem-
This expression also allows to estimate the value of ΘD. perature. Additionally, we didn’t observe any obvious,
We obtained the ΘD = 385(8) K which is very close to abnormal variation of hyperfine parameters around Tc.
the 408 K obtained from the analysis of the low temper- Two possibilities could lead to this result: the opening
ature specific heat capacity data and about 140 K less ofthesuperconductinggapdoesn’tbringvariationofthe
than the value estimated by temperature dependence of environmentatFe siteatall; theT ofthissystemis too
c
IS.Asimilardifferencewasfoundearlierinstudiesofe.g. low,andMössbauerspectroscopyis notsensitive enough
FeSe0.5Te0.5 and 57Fe- doped YBa2Cu3O6.8 compounds to detect the minute change.
[32, 34]. This discrepancy may be explained by the fact
the area reflects the averagemean-squaredisplacements,
whileISrelatedtothemean-squarevelocityoftheMöss-
bauer atom. Both quantities may respond in different
ACKNOWLEDGMENTS
ways to the lattice anharmonicities.
X. M. was supported in part by the China Scholar-
IV. CONCLUSIONS ship Council. The authors (H. P. and F. L.) grate-
fullyacknowledgethefinancialsupportfromtheNational
In summary, we performed detailed 57Fe Mössbauer Natural Science Foundation of China under grant No.
measurements on Lu2Fe3Si5 in the temperature range 11275086. Work at the Ames Laboratory (X. M., S. R.,
of 4.4 K to RT. The contributions from two Fe crystal- P. C. C. and S. L. B. ) was supported by the US De-
lographic sites can be well distinguished by Mössbauer partment of Energy, Basic Energy Sciences, Division of
spectra. The main contribution of EFG in this com- Materials Sciences and Engineering under Contract No.
pound comes from the lattice anisotropy and the first DE-AC02-07CH11358.
[1] R.L. Mössbauer, Z. Physik 151, 124 (1958). [14] V. O. Pelenovich, Y. Chen, R. Z. Xiao, S. Y. Liu, D. J.
[2] P. P. Craig, R. P. Taylor, and D. E. Nagle, Nuovo Ci- Fu, and Z.T. Yang,HyperfineInteract. 219, 1(2013).
mento22, 402 (1961). [15] J. K. M. Lindén, E. L. Rautama, M. Karppinen, and
[3] Y.Wu,S.Pradhan, andP.Boolchand,Phys.Rev.Lett. H. Yamauchi,HyperfineInteract. 208, 133 (2012).
67, 3184 (1991). [16] J. Munevar, H. Micklitz, J. Aguero, G. Tan, C. Zhang,
[4] T.Sinnemann,R.Job, andM.Rosenberg,Phys.Rev.B P. Dai, and E. Baggio-Saitovitch, Phys. Rev. B 88,
45, 4941 (1992). 184514 (2013).
[5] P. Boolchand, R. N. Enzweiler, I. Zitkovsky, J. Wells, [17] H. F. Branun,Phys. Lett. 75A, 386 (1980).
W. Bresser, D. McDaniel, R. L. Meng, P. H. Hor, and [18] H. Braun, C. Segre, and F. Acker, J. Magn. Magn.
C. W. Chu,Phys. Rev.B 37, 3766 (1998). Mater. 26, 117 (1981).
[6] C. M. Lin and S. T. Lin, J. Phys.: Condens. Matter 5, [19] J. Ge, J. Gutierrez, B. Raes, T. Watanabe, J. Koshio,
L247 (1993). and V.Moshchalkov, Physica C 478, 5 (2012).
[7] C.M.Lin,C.Y.Shyn, andS.T.Lin,J.Phys.: Condens. [20] B. H.Toby,J. Appl.Cryst. 34, 210 (2001).
Matter 6, L191 (1994). [21] Z. Klencsar, MossWinn 3.0i xp manual (1995).
[8] P.Boolchand,R.N.Enzweiler,I.Zitkovsky,R.L.Meng, [22] Y.Nakajima,T.Nakagawa,T.Tamegai, andH.Harima,
P. H. Hor, C. W. Chu, and C. Y. Huang, Solid State Phys. Rev.Lett. 100, 157001 (2008).
Commun. 63, 521 (1987). [23] J. D.Cashion, G.K.Shenoy,D.Niarchos, P.J. Viccaro,
[9] E.Kuzmann,M.Mair,Z.Klencsár,A.Vértes,Z.Homon- A. T. Aldred, and C. M. Falco, J. Appl. Phys 52, 2180
nay, and G. Gritzner, Physica C 319, 12 (1999). (1981).
[10] I.Nowik and I. Felner, Physica C 469, 485 (2009). [24] J. Cashion, G. Shenoy, D. Niarchos, P. Viccaro, and
[11] P. Wang, Z. M. Stadnik, J. Żukrowski, A. Thaler, S. L. C. M. Falco, Phys.Lett. 79A, 454 (1980).
Bud’ko, and P. C. Canfield, Phys. Rev. B 84, 024509 [25] M.Umarji,D.Noakes,P.Viccaro,G.Shenoy, andA.Al-
(2011). dred, J. Magn. Magn. Mater. 36, 61 (1983).
[12] X.-M.Ma, J.-M. Wan,J.-P. Liu, H.Pang, and F.-S. Li, [26] C. Blaauw, F. van der Woude, and G. A. Sawatzky, J.
EPL 105, 37002 (2014). Phys. C: Solid State Phys.6, 2371 (1973).
[13] G.Long,M.DeMarco,M.Chudyk,J.Steiner,D.Coffey, [27] P. Blaha, K. Schwarz, P. Sorantin, and S. B. Trickey,
H.Zeng,Y.K.Li,G.H.Cao, andZ.A.Xu,Phys.Rev. Comput. Phys. Commun. 59, 399 (1990).
B 84, 064423 (2011).
8
[28] P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, and [31] H. Pang, Y. Fang, and F. Li, Hyperfine Interact. 199,
J. Luitz, WIEN2K (K. Schwarz, Technische Univer- 387 (2011).
sitÂąÂğatWien,Austria),2001[http://www.wien2k.at]. [32] J. Lindén, J.-P. Libäck, M. Karppinen, E.-L. Rautama,
[29] J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. andH.Yamauchi,SolidStateCommun.151,130(2011).
Lett.77, 3865 (1996). [33] T. Sinnermann, R. Job, and M. Rosenberg, Phys. Rev.
[30] Z. Li, Y. Fang, X. Ma, H. Pang, and F. Li, Phys. Rev. B 45, 4941 (1992).
B 84, 134509 (2011). [34] V. Cherepanov, M. Chuev, E. Tsymbal, C. Sauer,
W. Zinn, S. Ivanov, and V. Zhurov, Solid State Com-
mun.93, 921 (1995).
