Control of Correlations in Sr4V2O6Fe2As2 by Chemical Stoichiometry
Athena S. Sefat, David J. Singh, V. Ovidiu Garlea, Yuri L. Zuev, Michael A. McGuire, Lindsay VanBebber, Brian C. Sales
CControl of Correlations in Sr V O Fe As by Chemical Stoichiometry Athena S. Sefat, David J. Singh, V. Ovidiu Garlea, Yuri L. Zuev, Michael A. McGuire, Lindsay VanBebber, and Brian C. Sales Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, TN, 37831, USA Department of Physics, University of Tennessee, Knoxville, TN, 37996, USA Department of Materials Science and Engineering, University of Tennessee, Knoxville, TN, 37996
We show using a combination of powder X-ray and neutron diffraction, first principles calcula-tions, temperature- and field-dependent magnetization, heat capacity and resistivity data that thesuperconducting behavior of ‘Sr V O Fe As ’ is dependent on synthesis conditions, particularly,heating profiles result in unintentional chemical doping. This compound can be tuned from a statein which the vanadium electrons are itinerant with a high electronic density of states, to a statewhere the vanadium-oxide layers are insulating and presumably antiferromagnetic. PACS numbers:
I. INTRODUCTION
Since the discovery of high-temperature superconduc-tivity in iron-based materials, five families of materi-als have been established. The ones about which mostis known are the rare earth-based R FeAsO materials,the alkaline-earth-based A Fe As compounds and themonochalcogenides Fe(Se,Te). These are known as the1111, 122, and 11 families, respectively. The parents (i.e.the undoped stoichiometric compounds) are metallic, andoften exhibit a spin-density-wave (SDW) type magnetictransition. Superconductivity emerges when the SDWis suppressed, typically via chemical substitution at ei-ther the R , or A or Fe-crystallographic site. The two other families of iron-based superconductorsare those of alkali-metal-based A FeAs and oxygen-basedSr T O Fe P n ( T = transition metals; P n = pnic-togens), the so-called 111 and 42622, respectively.
In 111 compounds, off-stoichiometry on the alkali-metalsite or deliberate chemical doping yields superconduc-tivity.
In contrast to most of the other compounds,the undoped parents of 42622 family are reported togive superconductivity, with transition temperatures of T C = 17 K for Sr Sc O Fe P and T C = 37 K forSr V O Fe As . Following those experimental reports,several theoretical papers have attempted to explainthe causes of such a high T C in Sr V O Fe As ; one group reported not being able to reproduce super-conductivity. The present work is motivated by theexperimental and theoretical controversies surroundingSr V O Fe As . Specifically, why does a stoichiomet-ric sample yield superconductivity? Is there a structuraltransition? Is there any type of magnetism or correlatedelectron behavior in the Sr V O layer? Related to thiswe note that vanadium oxides with V in an octahedralenvironment show an extraordinary variety of behaviorsrelated to the interplay of strong correlation effects, or-bital orderings, magnetism, and itinerancy. In particular,one can tune through metal-insulators transitions. Theseare often accompanied by strong changes in magneticorder and lattice structure. Sr V O Fe As combines perovskite-like, octahedrally-coordinated V oxide lay-ers and FeAs-superconducting layers, and therefore po-tentially offers a window into the interplay between acorrelated 3 d oxide and a high- T C iron based supercon-ductor.Like other families of Fe-based superconductors,Sr V O Fe As has a quasi-two-dimensional layeredcrystal structure, with iron in tetrahedral coordinationwith arsenic. The tetragonal symmetry of Sr FeO CuS-type ( P d -spacing ( ≈ . T C = 37K gives the lattice parameters a = 3 . c =15 . while the report of nonsuperconductivitygives a = 3 . c = 15 . Based onthis, one may infer that the presence or absence of super-conductivity depends on the slight changes in the struc-ture, due to unintended chemical doping. In iron-basedsuperconductors, small structural effects have been dis-cussed in relation to T C ; they include the structures ofa regular FeAs tetrahedron (109.47 ◦ ), and the specificpnictogen height from the Fe-plane ( h Pn ≈ .
37 ˚A).
Here we report six sets of polycrystalline syntheseswith different conditions, as well as physical property re-sults, in our attempts to produce ‘Sr V O Fe As .’ Eachsynthesis method gives different physical properties be-havior. We report an overview of syntheses followed byresults on the structure, thermodynamic and transportproperties, and first-principles calculations. We concludethat the parent (stoichiometric) 42622 may be magneticand is not superconducting, that unintentional chemical-doping (due to impurities) gives rise to superconductiv-ity, that the T C value is connected with the c -lattice pa-rameters which change with stoichiometry, and that thereis no structural transition for the superconducting sam-ples. We also note that the complex behavior of 42622 isconnected with the interplay between the V-O systemand the FeAs layers. Depending on composition, thematerial may cross from a non-superconducting, near-ferromagnetic itinerant V state that is characterizedby a much higher density-of-states, to a superconductingstate based on the insulating V layers. a r X i v : . [ c ond - m a t . s up r- c on ] S e p II. EXPERIMENTAL
For the polycrystalline Sr V O Fe As sample syn-thesis, ref. describes mixing stoichiometric amounts ofSrAs, V O , SrO, Fe and Sr, pressing them into a pelletform, and then heating them in a sealed silica tube at1150 ◦ C. Ref. describes the use of a different set of re-actants, namely V O , SrO , Sr, and FeAs, with heatingat 750 ◦ C followed by heating at 1150 ◦ C. Here, we reporta synthesis approach using another set of starting reac-tants, as we were unable to purify SrAs and avoided theuse of highly air-sensitive Sr powder. Initially, SrCO was decomposed to SrO by heating at 1325 ◦ C, and VAswas made through several heating steps in an evacuatedsilica tube (procedure: heat at 350 ◦ C, 600 ◦ C, and 800 ◦ C;regrind and heat at 800 ◦ C, and 1000 ◦ C; regrind and per-form final anneal at 1000 ◦ C). Our starting reactants wereSrO, VAs, Fe, and Fe O . They were mixed, pressed intopellets, and heated in evacuated silica tubes. Six pellets(labeled as Samples 1 through 6) were synthesized usingtwo different heating profiles (a and b described below).The cooling of all reactions involved quenching by turn-ing off the furnace.We find that the superconducting behavior is depen-dent on the heating profile. A description of each heat-ing profiles follows. Profile (a) involved heating to anintermediate temperature in the first synthesis step. Thepellet was then reground, repressed into a pellet, andheated to a high temperature in the second step. Usingthis outline, Samples 1, 2, and 3 were made. Sample 1was heated to 850 ◦ C (13 hrs), and then to 1150 ◦ C (24hrs). Sample 2 was heated to 550 ◦ C (5 hrs), and thento 1150 ◦ C (40 hrs). Sample 3 was heated to 850 ◦ C (13hrs), and then to 1100 ◦ C (24 hrs). Profile (b) involvedheating to a medium and a high temperature in the firstsynthesis step. The pellet was then reground, repressed,and heated to a high temperature in the next step(s).Samples 4, 5, and 6 were made using this profile. Sample4 was heated to 850 ◦ C (12 hrs) and 1100 ◦ C (24 hrs), thento 1100 ◦ C (24 hrs), then to 1150 ◦ C (24 hrs). Sample 5was heated to 850 ◦ C (15 hrs) and 1100 ◦ C (48 hrs), thento 1100 ◦ C (48 hrs). Sample 6 was heated to 850 ◦ C (15hrs) and 1150 ◦ C (40 hrs), then to 1150 ◦ C (12 hrs).The initial phase purity and structural identificationwere made via powder X-ray diffraction using PANalyt-ical X’pert PRO MPD. Lattice constants of all sampleswere determined at room temperature from LeBail refine-ments using X’Pert HighScore Plus (2.2e version). ForSample 4, powder x-ray diffraction data were collectedat 11 K with Cu K α radiation and an Oxford Phenixclosed-cycle cryostat.Powder neutron diffraction measurements were per-formed on two samples (2 and 3), each weighing ∼ We useda monochromatic beam with a wavelength of λ = 1 . He cryostat (2-300 K). The aluminum holderwas chosen to minimize the incoherent scattering and toimprove the detection of a potential magnetic scatter-ing. Additional experiments were carried out on Sam-ple 3 that was loaded in a vanadium holder. Such mea-surements were focused on investigating possible subtlechanges of the lattice across the superconducting phasetransition. Rietveld refinements were performed for eachset of data, at a given temperature, using the FULL-PROF program. DC magnetization was measured as a function of tem-perature using a Quantum Design Magnetic PropertyMeasurement System (MPMS). For a temperature sweepexperiment, the sample was zero-field cooled (ZFC) to 1.8K and data were collected by warming from 1.8 K in anapplied field. The sample was then field-cooled (FC) inthe applied field, and the measurement was repeated from1.8 K. The magnetic-susceptibility results are presentedper mole of formula unit (cm /mol), χ m .The temperature and field dependence of DC electri-cal resistance were measured using a Quantum DesignPhysical Property Measurement System (PPMS). Elec-trical contacts were placed on samples in standard four-probe geometry by using Pt wires and Dupont silverpaste. In order to exclude the longitudinal contributiondue to misalignment of Hall-voltage contacts, the Hallresistivity was derived from the antisymmetric part ofthe transverse resistivity under several magnetic field re-versal (range of 6 Tesla) at a given temperature, i.e. ρ H = [ ρ H (+ H ) − ρ H ( − H )] /
2. The Hall coefficient ( R H )was then evaluated from R H = ρ H /H , and a linear fitthrough the different H values. The temperature depen-dence of the heat capacity was also obtained using thePPMS, via the relaxation method.First principles calculations were done using the gen-eral potential linearized augmented planewave method as implemented in the WIEN2k code and also an in-house code. We used the Perdew Burke Ernzerhof (PBE)generalized gradient approximation and the local spindensity approximation. We also did PBE+U calculationswith the Coulomb parameter U − J = 6 eV, applied onlyto the vanadium. We based our calculations on the crys-tal structure of Zhu and coworkers. III. RESULTS AND DISCUSSION
Figure 1(a) shows the temperature dependence of mag-netic susceptibility ( χ m ) measured under ZFC and FCconditions at low magnetic fields (20 or 50 G) for Sam-ples 1 through 6. χ m ( T ) becomes negative for Samples1 to 5, indicating superconductivity. The ZFC and FCdivergence temperatures ( ≈ T C ) for these samples are 33K, 30 K, 25 K, 21 K, and 16 K, respectively. The highest T C value is consistent with the finding of the discoverypaper for Sr V O Fe As , which reports T C = 31 . χ ( T ) results. Samples 1 and 2 have the highest
FIG. 1: (a) Temperature dependence of molar susceptibility χ m in zero-field-cooled (ZFC) and field-cooled (FC) modes forsix differently prepared Sr V O Fe As samples (20 or 50 G).(b) Field dependence of magnetization for Samples 1 and 6at 2 K. T C values, as well as shielding and Meissner fractions.Assuming the theoretical density of ∼ . and a χ value of perfect diamagnetism, the shielding fractions are ∼
60% or 40%, respectively, and the Meissner fractionsnear the 10-15% range at 2 K. In Samples 2 and 3, despitethe high shielding fraction of ∼ ∼ J C of the order of few kA/cm . This lowvalue suggests that the grains are not coupled, and the persistent current circulates around each grain individu-ally, rather than the sample as a whole. Sample 6, with apositive value of χ m and weak divergence of ZFC-FC at ∼
15 K (Fig. 1a), shows a weak ferromagnetic signal at2 K (Fig. 1b), which exceeds the small superconductivediamagnetic signal. The magnetic hysteresis for Sample 6has a remanent magnetization value of M r = 0 . µ B /molof f.u. and a coercive field of 0.2 Tesla (inset of Fig. 1b).The room-temperature X-ray powder-diffraction dataidentified the main phase of all samples as Sr FeO CuS-structure-type ( P /nmm ). The diffraction data for Sam-ples 1 and 6 are presented in Figure 2. Sample 1, abulk superconductor with the highest T C , has the great-est number of impurity peaks (shown by asterisks). Incomparison, Sample 6, a non-bulk superconductor hasa cleaner diffraction pattern. The impurity phases areexpected to produce elemental deficiencies in the mainphase that would force the stoichiometry away from42622 in Sr V O Fe As . Using a Le Bail fit, the lat-tice constants of Sample 1 are refined as a = 3 . c = 15 . a = 3 . c = 15 . c -lattice parameterincreases with decreasing T C (sample number), while the a -lattice parameter does not show much change. Thechange of lattice parameters for the series of Samples isshown in inset of Figure 2. Superconductivity may besensitive to small structural changes and correlated withsmaller c -lattice parameter. Our results are consistentwith previous reports on Sr V O Fe As that yields a T C = 30 K for a sample with c = 15 . and nosuperconductivity for a sample with c = 15 . Reports find the arsenic height ( h As ) from the Fe-layer of ≈ .
38 ˚A is crucial for achieving maximum T C . ForSr V O Fe As , h As is reported as 1.42 ˚A and pressureexperiments have reported a higher T C of 46 K under 4GPa; these confirm the importance of smaller c -lattice FIG. 2: X-ray diffraction data for Samples 1 and 6 at roomtemperature. Asterisks denote impurity phases. Inset showsthe change of lattice parameters with Sample.
FIG. 3: The neutron powder diffraction data collected on thesuperconducting samples. (a) Measurements at various tem-peratures on Sample 2, indicating no significant change of thescattering; (b) Typical Rietveld refinement plot for Sample3, with vertical bars denoting the peak position of the mainphase (42622) and impurity phases (Sr VO , FeAs, SrO). parameter. In order to examine eventual phase transi-tions in the superconducting samples, neutron powderdiffraction measurements were performed on Samples 2and 3. A comparison of the data at four temperatures(2 K, 60 K, 150 K and 275 K) is shown in Figure 3(a).The data are shifted vertically so they can be better vi-sualized. No structural transformations were detectedor magnetic long range order within the instrument de-tection limit (0 . µ B /f.u. for an antiferromagnet). Also,powder x-ray diffraction data (not shown) confirm theabsence of structural transitions for Sample 4 down to11 K, giving c/a = 3.963 at 11 K and c/a = 3.986 at 300K. These values are comparable to neutron diffractiondata (Fig. 4).Additional neutron powder diffraction experimentswere carried out on Sample 3, for which measurementswere focused on investigating possible subtle changes ofthe lattice across T C . Rietveld refinements were per- FIG. 4: Temperature dependence of lattice parameters, a and c , as determined from the refinement of neutron powderdiffraction data for Sample 3. The ratio c/a follows a mono-tonic behavior and is well described by a cubic polynomialfunction. formed for each set of data (at a given temperature) usingthe FULLPROF program. We were able to detect a num-ber of impurity phases corresponding to Sr VO ( ∼ ∼ ∼ a = 7 . c = 6 . it has beensuggested that 7% vanadium-doping is responsible for thesuppression of superconductivity in 42622. We believethat such conclusions cannot be substantiated by neu-tron data since the scattering from vanadium is almostentirely incoherent. Also in this report, the c -latticeparameter of the super- and non-superconductive phaseswas found to remain unchanged, which is in disagreementto earlier reports and our findings, although this mayrelate to different synthesis conditions.Figure 4 displays the temperature dependence of therelative lattice parameters (∆ a/a and ∆ c/c ) of Sample3. One can observe a quite significant difference in thethermal behavior of the lattice spacing along the a and c crystallographic directions. The c parameter followsa monotonous trend with decreasing temperature whilethe a constant decreases less dramatically and flattensnear 100 K. The c/a ratio was further used to quantifythe evolution of the lattice distortion vs. temperature.This ratio is obviously dominated by and follows a sim-ilar trend to the c -lattice parameter. The trend is well FIG. 5: (a) Temperature dependence of normalized electri-cal resistivity measured upon cooling from room temperatureto 1.8 K, at zero field. Inset is the enlarged part of low-temperature data illustrating the shift of the onset of super-conducting transition temperature for 4. (b) Temperaturedependence of resistivity below 50 K, at zero and 10 T ap-plied magnetic fields. described by a cubic polynomial function, indicated bysolid line in Figure 4For all samples, the room temperature electrical re-sistivity ρ
300 K ranged from 3 to 8 mΩ · cm. Figure 5(a)shows the temperature-dependence of resistivity down to1.8 K, normalized to room temperature. Samples 1, 2and 4 show a plateau-like shape, in the 150 K to 200K region, while Samples 5 or 6 do not. This featurewas observed in the first report of the superconductingSr V O Fe As and was suggested to be the result of anincomplete suppression of an antiferromagnetic order. Our neutron diffraction data here shows no signs of mag-netic order for Sample 2 (Fig. 3), however, features arenoted in heat capacity (Fig. 7).Samples 1, 2 and 4 have zero resistance at low temper-atures, while Samples 5 and 6 do not. This suggests thatin the latter two samples, individual grains become su- perconducting on cooling, but do not form a continuouspath for current conduction. For Samples 1, 2 and 4, theonset transition temperatures (illustrated in inset of Fig.5a) for 90% of the normal-state value are T onset C ≈ ρ . = 0 . T C ≈ . Co . As (0.6 K), LaFe . Co . AsO (∆ T C ≈ and LaFeAsO . F . (∆ T C ≈ . The effects of 10 Tesla applied field on ρ ( T ) are shownin Figure 5(b). For Sample 1, the T C onset is only slightlyaffected and the resistance moves eventually to zero afterthat. The R = 0 temperature is suppressed by ∼
15 Kin sample 1, showing large vortex liquid regime togetherwith the effect of percolation, whereby the magnetic fieldquenches some of the grains that were available for con-duction in zero-field. This effect is pronounced most forsuch anisotropic materials. From the shift of T onset C wecan determine the slope dH C /dT near T C , although theresult depends on how we quantify the T onset C . Using thetwo tangents method, we obtain dH C /dT ∼ − , andthe use of Werthamer-Helfand-Hohenberg (WHH) equa-tion gives H C (0) = 0 . T C | dH C /dT | ≈
200 T. Thislarge H C (0) value together with the largest gap param-eter 2∆ /k B T C ≈ could be taken as an evidence that the para-magnetic (Chandrasekhar-Clogston) limit is exceeded in42622. On the other hand, if we define T onset C to presenta fixed 90% of the normal-state value ( ρ N ) at each field H , we obtain a smaller value of dH C /dT ≈ − H C (0) ≈
70 T, which does not exceed the paramagneticlimit.
FIG. 6: Temperature dependence of Hall coefficient measuredon three samples of Sr V O Fe As : Samples 1 and 4 are bulksuperconductors, and Sample 6 is not. FIG. 7: Temperature variation of heat capacity ( C p ) for vari-ous Sr V O Fe As (42622) samples. The data for BaFe As are shown in (a) for comparison with the 42622s. The en-larged C p region of 125 to 175 K for 42622s are shown in(b). The temperature-dependent Hall coefficients for Sam-ples 1, 4 and 6 are shown in Figure 6. The conduc-tion in these samples is either dominated by electron-or hole-like charge carriers. For each of the bulk super-conductors, Samples 1 or 4, R H ( T ) is mainly negativewith a sharp minimum at 38 K or 25 K, respectively.These results are consistent with that reported for su-perconducting Sr V O Fe As with R H = − × − to − . × − m /C for 50 to 160 K region. For the non-superconductor (Sample 6), R H is positive with a sharpmaximum near 7 K. It is interesting that R H changessign, depending on the samples. In each, the magnitudeof R H changes significantly across the measured temper-ature region (Fig. 5b). The temperature dependence of R H is probably caused by a multiband effect. In a multi-band model, R H is the weighted sum of the contributionsfrom each band and if the scattering rate of each band hasdifferent temperature dependence, then the weighted sum FIG. 8: (a) Temperature-dependent heat capacity measuredupon cooling and then warming for Samples 1 and 3. (b)Temperature-dependent electrical resistivity measured uponcooling and then warming for Sample 1. changes with temperature. The Fermi surfaces of iron-based superconductors consist of compensating holes andelectron sheets. As such, the mobilities of different bandscan have different temperature dependencies. Also, ther-mal expansion or slight difference in composition can pro-duce a redistribution of the carriers among the bands soas to create a strong temperature dependence of R H , aswell as a change in sign of R H . The inferred carrier con-centrations at 100 K are 1 . × cm − and 3 . × cm − for Samples 1 and 4, respectively, and 6 . × cm − for Sample 6, with the assumption of one-bandmodel.The temperature-dependent heat capacity, C p ( T ), wasmeasured upon cooling from 200 K to 1.8 K; the resultsare shown in Figure 7. There are only broad and barelyvisible features at T C . However, sharp and large anoma-lies are seen at ∼
150 K for Samples 1, 2, and 3, whileSample 4 gives a broad feature. Non-bulk superconduc-tors (Samples 5 and 6) show no peaks. Figure 7(a) also
FIG. 9: Temperature dependence of magnetic specific heatin the form of C M /T and entropy change ( S/R ) for (a)Sr V O Fe As Samples 1 and 4, and (b) BaFe As . plots the heat capacity data for BaFe As to show the rel-ative magnitude of the high temperature peaks comparedto 122’s structural/spin-density-wave feature at ∼
140 K.Cao and coworkers found a roughly 150 K transition inthe heat capacity in their “overdoped” 42622, and de-clared it as an antiferromagnetic transition, that was fol-lowed by a weak ferromagnetic transition at 55 K, andwith a T C = 24 K. Here we find that the 150 K transi-tion is sample dependent, correlated with the supercon-ductivity and that no transition is detected at ∼
55 K.Such sharp transitions in heat capacity are likely tobe intrinsic, although we suspected impurities. Search-ing for likely binary and ternary impurity phases, onlyV O is found to give a fairly sharp antiferromag-netic/structural transition in C p at ∼
150 K (upon cool-ing) that can be displaced to ∼
175 K (upon warming). We checked for such possible thermal hysteresis in Sam-ples 2 and 4, but did not find them (see Fig. 8a). C p wasfirst measured upon cooling from 200 K down to 120 K,then upon warming for the same region. There is no peak displacement in heat capacity for these samples, suggest-ing second-order transitions. Temperature-dependent re-sistivity was measured on cooling from room temperatureto 2 K, then on warming in the same region; this showsno feature (see Fig. 8b). Also, we have no evidence ofa crystalline V O impurity phase in x-ray and neutrondiffraction results. Considering that such phase could ex-ist amorphously, it would have to be the makeup of morethan 5% of each sample, in order to give the sharp fea-ture in C p ; this is unlikely. Therefore, we assume thatthese high-temperature peaks are an intrinsic feature ofthe bulk superconducting 42622 samples.In order to estimate the ‘magnetic’ heat capacity ( C M )for Samples 1 and 4, the non-magnetic contributions wereapproximated by Sample 6 (see Fig. 7b) and subtracted.The magnetic entropies ( S ) were estimated by integra-tion of C M /T versus T and are shown in Figure 9(a).The magnetic entropy associated with each peak is small(comparable to BaFe As ) and not recovered for the spinof 1/2 ground state for ln 2 = 0 .
69 (Fig. 9b). This smallentropy is expected to be associated with a small momentnot detectable by neutrons ( ∼ . µ B /vanadium).Figure 10(a) shows the temperature dependence ofheat capacity for ‘Sr V O Fe As ’ Samples in the formof C p /T versus T from 1.8 to 10 K. For Sample 1, thelinear fit of this region yields a residual Sommerfeld co-efficient of γ = 49 . − mol − (3.1 mJ K − molatom − ). Although this region is less linear for the restof samples, the γ value may be estimated and it con-sistently goes up with the Sample number (decrease in T C ). The γ value increases with sample purity (seeFig. 2), and is largest for Sample 6 at ∼
125 mJ K − mol − . A close examination of magnetic susceptibilityresults, displayed in Figure 10(b) and 1(b), suggest theonset of a ferromagnetic-like ordering. Such spin corre-lations are consistent with the enhanced γ and the de-struction of superconductivity. It is noteworthy that ourneutron diffraction measurements on Sample 6 would notbe able to detect a very small ferromagnetic moment, dueto the uncertainty in separating the magnetic from thenuclear scattering. However, the presence of an intrinsicferromagnetic order in our samples can in principle beprobed using polarized neutrons by applying a spin anal-ysis of the scattered neutrons; this will be the subjectof a future study. These γ values are much larger thanthose found in other Fe-based superconductors, for ex-ample LaFeAsO . F . ( γ = 4 . − mol − ) andBaFe . Co . As ( γ = 3 . − mol − ), or even theparent BaFe As ( γ = 6 . − mol − ).Our first principles calculations are similar to thosereported by others. In the LSDA and PBE calcula-tions, the vanadium layers contribute states at the Fermienergy and the ground state is ferromagnetic at leastwithin the V-O bilayers making up the perovskite partof the unit cell. It is possible that the ground state couldbe antiferromagnetic based on the c -axis stacking of fer-rromagnetic bi-layers between different unit cells or thatthere could be a spiral or more complicated antiferro- FIG. 10: For Sr V O Fe As samples, temperature depen-dence of (a) specific heat is shown in C p /T versus T formbelow 10 K, and also (b) magnetic susceptibility in 1000 G.The inset of (b) shows the non-linear behavior of inverse sus-ceptibility. magnetic state. We note that because of the metallicnature of FeAs planes, long-range magnetic interactions(e.g. RKKY type) should be present. Also, as notedby Mazin , the material is more two dimensional thanother Fe-based superconductors, and the energy differ-ences between different magnetic states are small, bothof which may depress and even entirely suppress the or-dering temperature. In any case, the ground state forSr V O Fe As is either ferromagnetic or close to ferro-magnetism at this level of theory.Of course, ferromagnetism is generally incompatiblewith singlet superconductivity, and even ferromagneticfluctuations are expected to be strongly pair breaking.Therefore even though, as discussed by Mazin, theFermi surface as related to the FeAs layers retains the fea-tures thought to be essential for superconductivity, thepresence of ferromagnetic or near ferromagnetic metal-lic V-O layers is expected to be highly antagonistic to superconductivity (note that scattering between the twosubsystems, even if they are quite distinct would be pairbreaking). In the PBE+U calculations, the V d -states aregapped away from the Fermi energy leaving the Fe bandsand the ground state as antiferromagnetic. For the non-spin-polarized state with the PBE functional, we obtaina total electronic density of states, N ( E F ) = 18 . − per unit cell, mostly from the vanadium. This high valuestrongly favors magnetism. It corresponds to a bare spe-cific heat coefficient, γ b = 43 mJ K − mol − (mole oftwo V atoms unit cell). With ferromagnetism we obtain N ( E F ) = 9 . − per cell, γ b = 22 . − mol − .For the PBE+U calculations, the vanadium carries mo-ments close to the expected S = 1 value, in particular1 . µ B within the vanadium LAPW sphere of radius2.05 Bohr. We find that N ( E F ) is essentially indepen-dent of the vanadium magnetic order, comes from the Feand has a lower value of 2.3 eV − per cell, correspondingto γ b = 10 . − mol − . The electronic structure hastwo electrons in the majority spin t g orbitals, which canaccommodate three electrons. This orbital degeneracy iswhat leads to orbital ordering in V compounds. Suchordering is possible in this compound, but the details willdepend on the exact crystal structure including atomicpositions, which is not known at present. Comparingthese values for the specific heat with the experimentaldata, we conclude that there are substantial enhance-ments above the bare values in both states. Nonethelesswe associate the high specific heat coefficients measuredin non-superconducting samples, with metallic vanadiumperhaps strongly enhanced by correlations, and the lowervalues in superconducting samples with non-metallic, in-sulating V. IV. CONCLUSIONS
The Sr V O Fe As compound is much harder to syn-thesize than other families and single crystals will be cru-cial for determining intrinsic behavior. We have foundthat the T C value is high for samples with most crys-talline impurities, and it decreases with increases in c -lattice parameter. There are no structural transitions inthese compounds, but there is some evidence of magneticorder. We believe that we can tune the V-O subsystemfrom a correlated metallic state to an insulating state,and that there is an interplay between the electronic be-havior of the V-O subsystem and the FeAs superconduc-tivity. This interplay is presumably magnetic in nature.It will be of considerable interest to elucidate the char-acter of magnetism associated with vanadium, both interms of ordering and spin fluctuations, especially viapolarized neutron scattering measurements. The resultssuggest that Sr V O Fe As is a compound that offersa window into the interplay of Fe-based superconductiv-ity and correlated oxide physics with tunability throughchemical stoichiometry. V. ACKNOWLEDGMENTS
Research at ORNL is sponsored by the Materials Sci-ences and Engineering Division, Office of Basic EnergySciences, US Department of Energy. We acknowledge discussions with D. K. Christen, V. Keppens and D. Man-drus for technical support. The work at High Flux Re-actor (ORNL) was sponsored by the Scientific User Fa-cilities Division, Office of Basic Sciences, U. S. DOE. Y. Kamihara, T. Watanabe, M. Hirano, H. Hosono, J. Am.Chem. Soc. 30 (2008), 3296. M. Rotter, M. Tegel, D. Johrendt, Phys. Rev. Lett. 101(2008), 107006. F. C. Hsu, J. Y. Luo, K. W. Yeh, T. K. Chen, T. W.Huang, P. M. Wu, Y. C. Lee, Y. L. Huang, Y. Y. Chu,D. C. Yan, M. K. Wu, Proc. National Academy Sci. 105(2009), 14262. C. Wang, L. Li, S. Chi, Z. Zhu, Z. Ren, Y. Li, Y. Wang,X. Lin, Y. Luo, S. Jiang, X. Xu, G. Cao, Z. Xu, Europhys.Lett. 83 (2008) 67006. Z. A. Ren, Wei L., J. Yang, W. Yi, X. L. Shen, Z. C. Li,G. C. Che, X. L. Dong, L. L. Sun, F. Zhou, Z. X. Zhao,Chin. Phys. Lett. 25 (2008), 2215. 16 Z. A. Ren, G. C. Che, X. L. Dong, J. Yang, W. Lu, W.Yi, X. L. Shen, Z. C. Li, L. L. Sun, F. Zhou, Z. X. Zhao,Europhys. Lett. 83 (2008), 17002. A. S. Sefat, A. Huq, M. A. McGuire, R. Y. Jin, B. C. Sales,D. Mandrus, L. M. D. Cranswick, P. W. Stephens, K. H.Stone, Phys. Rev. B 78 (2008), 104505. A. S. Sefat, R. Y. Jin, M. A. McGuire, B. C. Sales, D. J.Singh, D. Mandrus, Phys. Rev. Lett. 101 (2008), 117004. S. Sharma, A. Bharathi, S. Chandra, V. R. Reddy, S.Paulraj, A. T. Satya, V. S. Sastry, A. Gupta, C. S. Sundar,Phys. Rev. B 81 (2010), 174512. S. Jiang, H. Xing, G. Xuan, C. Wang, Z. Ren, C. Feng, J.Dai, Z. Xu, G. Cao, J. Phys.:Condens. Matter 21 (2009),382203. K. W. Yeh, T. W. Huang, Y. L. Huang, T. K. Chen, F.C. Hsu, P. M. Wu, Y. C. Lee, Y. Y. Chu, C. L. Chen, J.Y. Luo, D. C. Yan, M. K. Wu, Europhys. Lett. 84 (2008),37002. Y. Mizuguchi, F. Tomioka, S. Tsuda, T. Yamaguchi, Y.Takano, J. Phys. Soc. Japan 78 (2009), 074712. Y. Mizuguchi, F. Tomioka, S. Tsuda, T. Yamaguchi, Y.Takano, Appl. Phys. Lett. 94, 012503 (2009). M. J. Pitcher, D. R. Parker, P. Adamson, S. J. C. Herkel-rath, A. T. Boothroyd, R. M. Ibberson, M. Brunelli, S. J.Clarke, Chem. Commun. 45 (2008), 5918. H. Ogino, Y. Matsumura, Y. Katsura, K. Ushiyama, S.Horii, K. Kishio, J. Shimoyama, Supercond. Sci. Technol.22 (2009), 075008. X. Zhu, F. Han, G. Mu, P. Cheng, B. Shen, B. Zeng, H.H. Wen, Phys. Rev. B 79 (2009), 220512(R). X. C. Wang, Q. Q. Liu, Y. X. Lv, W. B. Gao, L. X. Yang,R. C. Yu, F. Y. Li, C. Q. Jin, Solid State Comm. 148 (2008), 538. D. R. Parker, M. J. P. Smith, T. Lancaster, A. J. Steele,I. Franke, P. J. Baker, F. L. Pratt, M. J. Pitcher, S. J.Blundell, S. J. Clarke, arXiv:0909.0417 (2009). T. L. Xia, J. B. He, D. M. Wang, G. F. Chen,arXiv:1001.3311 (2010). I. R. Shen, J. Supercond. Nov. Magn. 22 (2009), 613. K. W. Lee, W. E. Pickett, Europhys. Lett., 89 (2010),57008. I. I. Mazin, Phys. Rev. B. 81 (2010) 020507(R). A. Pal, J. Supercond. Nov. Magn. 22 (2009), 619. C. H. Lee, A. Iyo, H. Eisaki, H. Kito, M. T. Fernandes-Diaz, T. Ito, K. Kihou, H. Matsuhata, M. Braden, K. Ya-mada, J. Phys. Soc. Japan 77 (2008), 083704. Y. Mizuguchi, Y. Hara, K. Deguchi, S. Tsuda, T. Yam-aguchi, K. Takeda, H. Kotegawa, H. Tou, Y. Takano, Su-percond. Sci. Technol. 23 (2010), 054013. E. Z. Kuchinskii, I. A. Nekrasov, M. V. Sadovskii,arXiv:1004.0801 (2010). O. Garlea, B. C. Chakoumakos, S. A. Moore, G. B. Taylor,T. Chae, R. G. Maples, R. A. Riedel, G. W. Lynn, D. L.Selby, Appl. Phys. A 99 (2010), 531. J. Rodriguez-Carvajal, Physica B 192 (1993), 55. D. J. Singh and L. Nordstrom, Planewaves Pseudopo-tentials and the LAPW Method, 2nd Edition (Springer,Berlin, 2006). P. Blaha, K. Schwarz, G. Madsen, D. Kvasnicka, and J.Luitz, WIEN2k, An Augmented Planewave + Local Or-bitals Program for Calculating Crystal Properties (Tech.Univ., Wien, Austria, 2001). H. Kotegawa,T. Kawazoe, H. Tou, K. Murata, H. Ogino,K. Kishio, J. Shimoyam, J. Phys. Soc. Japan 78 (2009),123707. M. Tegel, T. Schmid, T. Sturzer, M. Egawa, Y. Su, A.Senyshyn, D. Johrendt, arXiv:1008.2687. A. S. Sefat, M. A. McGuire, B. C. Sales, R. Jin, J. Y.Howe, D. Mandrus, Phys. Rev. B. 77 (2008), 174503. M. Hiraishi, R. Kadono, S. Takeshita, M. Miyazaki, A.Koda, H. Okabe, J. Akimitsu, J. Phys. Soc. Jpn. 78 (2009),023710. G. Cao, Z. Ma, C. Wang, Y. Sun, J. Bao, S. Jiang, Y.Luo, C. Feng, Y. Zhou, Z. Xie, F. Hu, S. Wei, I. Nowik, I.Felner, L. Zhang, Z. Xu, F. Zhang, arXiv:1007.3980.36