Combined effects of transition metal (Ni and Rh) substitution and annealing/quenching on physical properties of CaFe 2 As 2
Sheng Ran, Sergey L. Bud'ko, Warren E. Straszheim, Paul C. Canfield
aa r X i v : . [ c ond - m a t . s up r- c on ] J un Combined effects of transition metal (Ni and Rh) substitutionand annealing/quenching on physical properties of CaFe As S. Ran,
1, 2
S. L. Bud’ko,
1, 2
W. E. Straszheim, and P. C. Canfield
1, 2 Ames Laboratory, Iowa State University, Ames, Iowa 50011, USA Department of Physics and Astronomy,Iowa State University, Ames, Iowa 50011, USA (Dated: February 17, 2018)
Abstract
We performed systematic studies of the combined effects of annealing/quenching temperature( T A/Q ) and T = Ni, Rh substitution ( x ) on the physical properties of Ca(Fe − x T x ) As . Weconstructed two-dimensional, T A/Q - x phase diagrams for the low-temperature states for both sub-stitutions to map out the relations between ground states and compared them with that of Co-substitution. Ni-substitution, which brings one more extra electron per substituted atom andsuppresses the c -lattice parameter at roughly the same rate as Co-substitution, leads to a similarparameter range of antiferromagnetic/orthorhombic in the T A/Q - x space as that found for Co-substitution, but has the parameter range for superconductivity shrunk (roughly by a factor oftwo). This result is similar to what is found when Co- and Ni-substituted BaFe As are com-pared. On the other hand, Rh-substitution, which brings the same amount of extra electrons asdoes Co-substitution, but suppresses the c -lattice parameter more rapidly, has a different phasediagram. The collapsed tetragonal phase exists much more pervasively, to the exclusion of the nor-mal, paramagnetic, tetragonal phase. The range of antiferromagnetic/orthorhombic phase space isnoticeably reduced, and the superconducting region is substantially suppressed, essentially trun-cated by the collapsed tetragonal phase. In addition, we found that whereas for Co-substitutionthere was no difference between phase diagrams for samples annealed for one or seven days, for Ni-and Rh- substitutions a second, reversible, effect of annealing was revealed by seven-day anneals. PACS numbers: 74.70.Xa, 61.50.Ks, 75.30.Kz . INTRODUCTION Among the parent compounds of the Fe-based superconductors, CaFe As manifestsunique physical properties and has become a model system for understanding high- T c su-perconductivity in Fe-based superconductors. The magnetic and structural phase tran-sitions are strongly coupled and first order with hysteresis of several degrees as seen inthermodynamic, transport, and microscopic measurements.
Also, CaFe As is the mostpressure sensitive of the AFe As (A = Ba, Sr, Ca) and 1111 compounds with its anti-ferromagnetic/orthorhombic (AFM/ORTH) phase transition being initially suppressed byover 100 K per GPa and a then non-moment bearing, collapsed tetragonal (cT) phase beingstabilized by ∼ Previous work shows that the phase transition temperatures and even ground state ofCaFe As can be controlled and tuned by post-growth annealing and quenching of singlecrystal samples. Crystals of CaFe As grown from Sn-flux and quenched from 600 ◦ C man-ifest AFM/ORTH phase transition at 170 K. On the other hand, we found that crystalsgrown from FeAs-rich solutions, quenched from higher temperature (960 ◦ C) exhibit a tran-sition to the cT phase below 100 K at ambient pressure. Further, we found that, for theFeAs-flux grown samples, a process of post-growth annealing and quenching can be usedas an additional control parameter to tune the ground state of CaFe As systematically,suppressing the AFM/ORTH transition temperature from 170 K to ∼
100 K and then sta-bilizing an ambient pressure cT phase by changing T A/Q from 350 ◦ C to 800 and further to960 ◦ C. Time dependence of annealing effects was studied to establish annealing protocolsand the T - T A/Q phase diagram was constructed. It was found that the T - T A/Q and the T - P phase diagrams are similar, suggesting that the effect of annealing and quenching is simi-lar to that of pressure. Based on the TEM results, which reveal nano-scale, low strainprecipitates in the sample with T A/Q = 500 ◦ C and a strain field in the sample with T A/Q = 960 ◦ C, it is likely that the annealing and quenching process controls how a small excessof FeAs is brought in and out of the CaFe As matrix and, as a result, the amount of stressand strain built up in the samples, mimicing the modest pressures needed to stabilize thecT phase.Having mastered the control of FeAs grown CaFe As , we are able to tune the groundstate of this system with the combination of chemical substitution, annealing/quenching2nd application of hydrostatic pressure. The combined effects of Co-substitution andannealing/quenching on the physical properties of CaFe As were studied and a 3D phasediagram, with Co concentration, x , and annealing/quenching temperature, T A/Q , as two in-dependent control parameters, was constructed. At ambient pressure, the Ca(Fe − x Co x ) As system offers ready access to the salient low-temperature states associated with Fe-based superconductors: antiferromagnetic/orthorhombic (AFM/ORTH), superconduct-ing/paramagnetic/tetragonal (SC/PM/T), non superconducting/paramagnetic/tetragonal(N/PM/T) and non-moment bearing/collapsed tetragonal (cT). In a similar manner, for x = 0.028, very modest, hydrostatic pressure ( P <
260 MPa) can change the low-temperatureground state from AFM/ORTH to SC/PM/T, to a cT phase. Detailed comparisons of the P and T A/Q dependence of the Ca(Fe − x Co x ) As phase diagrams further supported thesimilarity of the effects of T A/Q and P on these compounds. For example, for x = 0.028, ascaling of ∆ T A/Q = 100 ◦ C being equivalent to ∆ P = 85 MPa allowed for the T - T A/Q andthe T - P phase diagrams to fall on a single manifold. In the case of BaFe As system, comparison of the phase diagrams of various transi-tion metal substitutions reveals that while the suppression of the structural/magnetic phasetransition scales, roughly, with impurity concentration, x , the superconductivity is rathercontrolled by extra electron count, e . Steric effect seems not to play any important role indetermining the phase diagram, with Co- (Ni-) and Rh- (Pd-) substitution having excep-tionally similar effect, especially on the superconducting dome on the overdoped side.In order to compare the phase diagrams of various transition metal substitutions inCaFe As , we expand the exploration of transition metal substitution to Ni and Rh. Com-pared with Co-substitution, Ni-substitution brings one more extra electron per substitutedatom, while suppressing c -lattice parameter in a very similar manner. On the other hand,Rh-substitution brings nominally the same amount of extra electrons as Co-substitution,although from a 4d-shell rather than a 3d-shell, while suppressing the c -lattice parametermuch more rapidly. Therefore, comparing Co-substitution with Ni- and Rh-substitution willpotentially help us understand the changes of physical properties of CaFe As system causedby (i) band filling and (ii) steric effect. As we will show, this is more complicated than inthe case of Ba(Fe − x T x ) As , not only due to the existence of one more control parameter, T A/Q , but also because CaFe As is much more sensitive to the pressure, and therefore tothe steric effect. 3ue to the fact that two independent control parameters ( T A/Q and x ) define this phasespace, large amounts of temperature dependent data were collected and used to assemblethe various phase diagrams. In the main body of this paper only selected sets of data willbe presented and the rest of data will be presented in the Appendix. II. EXPERIMENTAL METHODS
Single crystals of Ca(Fe − x TM x ) As were grown out of FeAs-flux, using conventionalhigh-temperature solution growth techniques. The growth protocol and the post-growth thermal treatments for both Ni- and Rh-substituted CaFe As are the same as forCo-substitution. In the process of decanting off the excess flux, the samples were essen-tially quenched from 960 ◦ C to room temperature, which, according to our previous study,causes strain inside the samples, leading to behavior different from Sn grown samples.
These samples will be referred to as T A/Q = 960 ◦ C (as grown) samples. post-growth, ther-mal treatments of samples involve annealing samples at a certain temperature ranging from350 ◦ C to 800 ◦ C and subsequently quenching from this annealing temperature to room tem-perature. These samples will be identified as T A/Q = 350 ◦ C to T A/Q = 800 ◦ C. The datapresented in the first part of the text were all collected on the samples annealed for 24 hours.As will be discussed in the second part of this paper, for Ni- and Rh-substitution, longerannealing times appear to further introduce a second, controllable process. Details aboutthe annealing and quenching technique can be found in references. .Elemental analysis was performed on each of these batches using wavelength-dispersivex-ray spectroscopy (WDS) in the electron probe microanalyzer of a JEOL JXA-8200 elec-tron microprobe. Since the properties of a given sample are found to be determined byboth the transition metal substitution level, x , and the post-growth annealing/quenchingtemperature, T A/Q , samples are fully identified by providing both of these parameters.Diffraction measurements on the platelike samples were performed at room temperatureusing a Rigaku Miniflex diffractometer with Cu K α radiation. Only (00l) peaks are observedfrom which the values of the c -lattice parameter are inferred. Standard powder x-ray diffrac-tion was not attempted because grinding of CaFe As leads to severe peak broadening andeven changes in physical properties, as has been outlined in our previous work. Temperature dependent DC magnetization measurements were made in Quantum Design4agnetic Property Measurement Systems (MPMS). The in-plane, temperature dependent,electrical resistance measurements were performed in Quantum Design Physical PropertyMeasurement Systems (PPMS) or in Quantum Design MPMS systems operated in externaldevice control (EDC) mode, in conjunction with Linear Research LR700 four-probe ACresistance bridges (f =16Hz, I = 1mA). The electrical contacts were placed on the samplesin standard 4-probe geometry, using Pt wires attached to a sample surface with EpotekH20E silver epoxy.In order to infer phase diagrams from these thermodynamic and transport data, we needto introduce criterion for the determination of the salient transition temperatures.
TheAFM/ORTH phase transition (when present) appears as a single (i.e., not split), sharpfeature which is clearly identifiable in both resistance and magnetization. Figure 1 showsthe susceptibility and resistance, as well as their temperature derivatives (insets), for a x = 0.006/ T A/Q = 400 ◦ C Rh-substituted sample. Clear features, including a sharp drop insusceptibility and a sharp jump in the resistance, occur upon cooling through the transi-tion temperature. The transition temperature is even more clearly seen in the d ( M/H ) /dT and dR/dT data. For the superconducting transition, we only used an onset criterion formagnetic susceptibility (the temperature at which the maximum slope of the susceptibilityextrapolates to the normal state susceptibility) to determine T c . This criteria for T c ispresented in Fig.2a, with an example of a x = 0.023/ T A/Q = 400 ◦ C Rh-substituted sam-ple. Sometimes an offset criterion for resistance (the temperature at which the maximumslope of the resistance extrapolates to zero resistance) is also used in literature to determine T c . However, this leads to substantially higher T c than ones inferred from magnetic sus-ceptibility in CaFe As which, given its profound pressure and strain sensitivities, is proneto filamentary superconductivity (Fig. 2b). Given that we do observe diamagnetism withzero field cooling (ZFC) susceptibility reaching 1/4 π , we choose to err on the side of bulksuperconductivity rather than minority phase of filamentary superconductivity. The cTphase is induced by higher T A/Q . When the cT phase transition occurs, it often leads tocracks in the resistance bar and loss of data below the transition temperature (in case theresistance bar survives upon cooling through the transition, resistance data shows downwardjump and hysteresis of up to around 15 K), which is an unique fingerprint of the cT phasetransition and helps us to distinguish it from AFM/ORTH phase transition. On the otherhand, loss of data below the transition makes it difficult to extract an unambiguous value of5he transition temperature from R ( T ) data. Therefore only susceptibility data were used todetermine the transition temperature, T cT . Figure 3 shows the susceptibility data for twodifferent samples. The peak in derivative of the susceptibility was employed to determine T cT . Note that the peak in derivative becomes significantly broadened for high concentra-tions, as shown for the x = 0.049/ T A/Q = 400 ◦ C Rh-substituted sample. We capture thebroadness of the transition by including error bars, which were defined here as the full widthat half maximum of the peaks in derivatives of the susceptibility.
Rh x = 0.006/T
A/Q = 400(cid:176)C M H ( - e m u / m o l e ) (a) (b) R / R ( K ) T (K)
120 130 140 150 160 1706.57.07.5 M H ( - e m u / m o l e ) T (K)
120 130 140 150 160 1700.60.8 R / R ( K ) T (K) -0.15-0.10-0.050.00 d [ R / R ( K ) ]/ d T ( / K ) d ( M H ) / d T ( - e m u / m o l e K ) FIG. 1: (Color online) Criteria used to determine the transition temperatures of the AFM/ORTHphase transition. The data close to the transition are presented in the insets, together with thederivatives. Inferred transition temperatures are indicated by vertical arrows.
III. RESULTS AND DISCUSSIONCompositional and structural determination
A summary of the WDS measurement data for both Ni- and Rh-substituted compoundsis presented in Fig. 4. Data for the Ca(Fe − x Co x ) As series are also presented for com-parison. The nominal concentration versus actual concentration data for all three series can6 (a) Rh x = 0.023/T
A/Q = 400(cid:176)C M / H ( / ) (b) R / R ( K ) T (K) FIG. 2: (Color online) Criteria used to determine the transition temperatures of the superconduct-ing phase transition. Inferred transition temperatures are indicated by arrows. As discussed in thetext, use of resistivity data can lead to artificially high T c values due to strain-induced filamentarysuperconductivity. be fitted very well with straight lines, indicating a linear correlation between the measuredconcentration and the nominal concentration for these relatively low ( x < − x Co x ) As and Ca(Fe − x Ni x ) As are close to 1 (0.96 ± ± ± − x Rh x ) As . Theerror bars are taken as twice of the standard deviation determined from the 12 WDS mea-surements on each sample, and are no more than 0.003, demonstrating relative homogeneityof the substituted samples studied here. In the following, the average experimentally de-termined x values, x = x W DS , will be used to identify all the compounds rather than thenominal concentration, x nominal .Figure 5 presents the c -lattice parameters for the T A/Q = 960 ◦ C samples, as well asfor the T A/Q = 400 ◦ C samples, for both Ca(Fe − x Ni x ) As and Ca(Fe − x Rh x ) As series,determined via diffraction from the platelike samples using (002) and (008) peaks. Data forthe Ca(Fe − x Co x ) As series obtained in a similar way are also presented for comparison.7 (a) Rh x = 0.028/T
A/Q = 400(cid:176)C M H ( - e m u / m o l e ) Rh x = 0.049/T
A/Q = 400(cid:176)C (b) M H ( - e m u / m o l e ) T (K)
50 60 70 80 90246 M H ( - e m u / m o l e ) T (K)
50 100 150 200 25068 M H ( - e m u / m o l e ) T (K) d ( M H ) / d T ( - e m u / m o l e K ) -0.050.000.05 d ( M / H ) / d T ( - e m u / m o l e K ) FIG. 3: (Color online) Criteria used to determine the transition temperatures of the cT phasetransition shown for two different samples. The data close to the transition are presented in theinsets, together with the derivatives. Inferred transition temperatures are indicated by verticalarrows.
It can be seen that in case of both the T A/Q = 960 ◦ C samples and the T A/Q = 400 ◦ Csamples the c -lattice parameter is suppressed by all three transition metal substitutions.Whereas Ni-substitution suppresses the c -lattice parameter at roughly the same rate as Co-substitution, Rh-substitution suppresses the c -lattice parameter roughly twice as fast. Thisis similar to what has been seen for BaFe As . However, in BaFe As this difference in thesuppression of the c -lattice parameter does not seem to matter much in terms of its effect onthe T - x phase diagrams, i.e. Co- and Rh-substitutions being virtually identical but differingfrom Ni- and Pd-substitutions (each with an extra conduction electron). Considering thatthe physical properties of CaFe As are much more sensitive to the stress and strain thanare those of BaFe As , the large suppression of the lattice parameter in Ca(Fe − x Rh x ) As series may have a much more dramatic effect than in the case of Ba(Fe − x Rh x ) As .8 .00 0.02 0.04 0.06 0.08 0.100.000.020.040.06 Co Ni Rh x W D S x nominal FIG. 4: (Color online) Measured Co, Ni and Rh concentration vs nominal Co, Ni and Rh concen-tration for the Ca(Fe − x T x ) As series. Data of Co-substitution are taken from Ref. Ca(Fe − x Ni x ) As Figure 6 presents the temperature dependent magnetic susceptibility and resistance datafrom Ca(Fe − x Ni x ) As samples with T A/Q = 400 ◦ C. The x = 0, parent compound, showsAFM/ORTH phase transition at around 166 K as indicated by the sharp drop in suscepti-bility and upward jump in resistance upon cooling. The anomalies in both susceptibility andresistance are suppressed with increasing Ni substitution level, down to 55 K, at x = 0.023.For higher Ni concentrations, the AFM/ORTH phase transition is suppressed completelyand the SC/PM/T phase is stabilized. At x = 0.025, low field susceptibility shows that thescreening is around 60% of 1/4 π at 2 K as shown in Fig. 6b. At x = 0.027, the screeningincreases to 100% of 1/4 π . Above x = 0.027, increasing Ni concentration suppresses T c andthe screening. For x > T A/Q (Fig. 6d). TheAFM/ORTH phase transition is suppressed with initial Ni-substitution and the phase lineterminates at around x = 0.025, where the SC/PM/T phase emerges. T c is suppressed9 .00 0.02 0.04 0.0611.3511.4011.4511.5011.5511.6011.65 (a) Co Ni Rh c ( ¯ ) x T A/Q = 960(cid:176)C T A/Q = 400(cid:176)C
Co Ni Rh c ( ¯ ) x (b) FIG. 5: (Color online) Room temperature c -lattice parameter for Ca(Fe − x Ni x ) As andCa(Fe − x Rh x ) As series, determined via diffraction from plate like samples, as described in theExperimental Methods section, as a function of measured Ni/Rh concentration, x for (a) T A/Q = 960 ◦ C samples and (b) T A/Q = 400 ◦ C samples. For comparison, data for Ca(Fe − x Co x ) As samples from Ref. 12 are also presented. by further increasing Ni concentration. Bulk superconductivity, as indicated by low fieldsusceptibility, is suppressed completely for x > .00 0.02 0.04 0.06050100150 N/PM/T T S /T N -M T S /T N -R T c -M N/PM/T T r an s i t i on T ( K ) x (d) AFM/ORTH
SC/PM/T R / R ( K ) T (K) (c) (b) H = 0.01T cZFC M / H ( / ) T (K) Ni - T A/Q = 400(cid:176)C M H ( - e m u / m o l e ) T (K) H = 1T ll c (a) x =
FIG. 6: (Color online) Temperature dependent (a) magnetic susceptibility with field applied parallelto the c axis, (b) low-field magnetic susceptibility measured upon zero field cooling (ZFC) witha field of 0.01 T applied perpendicular to the c axis, (c) normalized electrical resistance, and (d)phase diagram of transition temperature T vs Ni concentration x of Ca(Fe − x Ni x ) As samples with T A/Q = 400 ◦ C. Susceptibility data in (a) have been offset from each other by an integer multiple of1 × − emu/mole for clarity. Yellow symbols indicate N/PM/T state with no transition observeddown to the base temperature. of the magnetic and structure phase transition is observed.In a similar manner, we constructed T - x phase diagrams for other T A/Q values as presentedin Fig. 7a and b (see Appendix for corresponding magnetic susceptibility and resistancedata). T A/Q = 500 ◦ C leads to similar phase diagram, but, both the AFM/ORTH and theSC/PM/T phase regions are reduced. This is consistent with the fact that increasing T A/Q has the same effect as increasing pressure as shown for pure and Co-substituted compoundsin our previous work.
For T A/Q = 960 ◦ C, CaFe As transforms into a cT state at low11emperature. As x is increased in the Ca(Fe − x Ni x ) As series the transition temperatureof this cT phase is gradually suppressed. This is in contrast to the Ca(Fe − x Co x ) As series,where this phase transition occurs at roughly the same temperature throughout the wholesubstitution level in our study range. Around x = 0.043, the signature of transition isbroadened and, as discussed in the experimental section, the large error bar is meant torepresent this. SC/PM/T N/PM/T (a) Ni - T A/Q = 500(cid:176)C T r an s i t i on T ( K ) x AFM/ORTH T S /T N -M T S /T N -R T c -M T cT -M N/PM/T
400 600 800 1000050100150
N/PM/T cTSC/PM/TAFM/ORTH (d) T r an s i t i on T ( K ) T A/Q ((cid:176)C) Ni - x = 0.026
400 600 800 1000050100150
N/PM/T cTAFM/ORTH (c) Ni - x = 0.021 T r an s i t i on T ( K ) T A/Q ((cid:176)C) cT N/PM/T (b) T r an s i t i on T ( K ) x Ni - T A/Q = 960(cid:176)C
FIG. 7: (Color online) phase diagram of (a) transition temperature T vs Ni concentration x ofCa(Fe − x Ni x ) As samples with T A/Q = 500 ◦ C, (b) transition temperature T vs Ni concentration x of Ca(Fe − x Ni x ) As samples with T A/Q = 960 ◦ C, (c) transition temperature T vs anneal-ing/quenching temperature T A/Q of Ca(Fe − x Ni x ) As samples with Ni concentration x = 0.021,and (d) transition temperature T vs annealing/quenching temperature T A/Q of Ca(Fe − x Ni x ) As samples with Ni concentration x = 0.026. Yellow symbols indicate N/PM/T state with no transitionobserved down to the base temperature.
12n order to systematically study the effect of the varying T A/Q for a given Ni substitu-tion level, we studied x = 0.021 and x = 0.026 samples for 350 ◦ C T A/Q ◦ C. Thecorresponding phase diagrams are presented in Fig. 7c and d. For both x = 0.021 and x =0.026, the ground state of the Ca(Fe − x Ni x ) As series is AFM/ORTH phase for low T A/Q ( ◦ C for x = 0.021 and ◦ C for x = 0.026) and cT phase for high T A/Q ( > ◦ C).For intermediate values of T A/Q , no bulk superconductivity (i.e., with significant screening)is observed for x = 0.021, whereas for x = 0.026, bulk superconductivity with screening ofmore than 70% of 1/4 π at 2 K is observed for T A/Q = 400 ◦ C.Since in this work we mainly focus on mapping out the relationship between possible lowtemperature states for various combinations of substitution level and annealing/quenchingtemperature, we can construct a 2D phase diagram, with transition metal concentration x and annealing/quenching temperature T A/Q as two independent variables, and mark theground state with different symbols (and colors online). This phase diagram is essentially aprojection of the 3D phase diagram (as in Ref. ) onto the plane of base temperature. Basedon the magnetic susceptibility and resistance data, we assembled a 2D phase diagram forNi-substitution and compare it with that of Co-substitution, as shown in Fig. 8.As seen for the Ca(Fe − x Co x ) As system, the Ca(Fe − x Ni x ) As system also pos-sesses the same salient low temperature states associated with Fe-based superconductors:AFM/ORTH, SC/PM/T, N/PM/T and cT. The AFM/ORTH region found for both Ni- andCo-substitution span essentially the same parameter space, whereas the SC/PM/T regionfor Ni-substitution is significantly reduced compared with Co-substitution, with maximum x value that supports SC/PM/T being much smaller than that for Co-substitution (for T A/Q = 350 ◦ , the critical substitution level is roughly 3.5% for Ni and 5.5% for Co). Thisis consistent with what was found for Co- and Ni-substituted BaFe As , where anti-ferromagnetism seems to be primarily controlled by the impurity concentration x whereassuperconductivity was more strongly influenced by extra electron count, e . In addition, thecT phase region for Ni-substituted CaFe As is also reduced. The cT phase is only stabilizedfor T A/Q > ◦ C as opposed to T A/Q > ◦ C for Co-substitution.13
IG. 8: (Color online) 2D phase diagrams, with transition metal concentration x and anneal-ing/quenching temperature T A/Q as two independent variables, for (a) Ca(Fe − x Co x ) As , (b)Ca(Fe − x Ni x ) As , and (c) Ca(Fe − x Rh x ) As . The red area delineates the conditions that leadto AFM/ORTH phase as ground state. The green area delineates the conditions that lead toSC/PM/T phase as ground state. The yellow area delineates the conditions that lead to N/PM/Tphase as ground state. The blue area delineates the conditions that lead to cT phase as groundstate. Ca(Fe − x Rh x ) As Rh-substitution brings the same nominal amount of extra electrons as Co-substitutiondoes and, despite the generic difference between 3d-shell and 4d-shell electrons as well as amuch more rapid change in the c lattice parameter. In the case of Ba(Fe − x Rh x ) As , Rh-substitution leads to a virtually identical T - x phase diagram as found for Ba(Fe − x Co x ) As .Given that CaFe As is much more sensitive to the pressure and strain than BaFe As , the14ifferent steric effects may well lead to differences in the T A/Q - x phase diagrams in the caseCa(Fe − x Co x ) As of Ca(Fe − x Rh x ) As .Figures 9a to c present the magnetic susceptibility and resistance data forCa(Fe − x Rh x ) As compounds with T A/Q = 400 ◦ . Rh-substitution initially suppresses theAFM/ORTH transition to below 50 K by x = 0.02. Bulk superconductivity is observed in asmall region of x value, as shown by screening in low field susceptibility (9b). Unlike the casesof Co- or Ni-substitution, both of which which have a region of T A/Q - x values that lead to aN/PM/T ground state without bulk superconductivity (8a and b), Rh-substitution stabilizesthe cT state much more rapidly, precluding any N/PM/T phase and abruptly terminating itsSC/PM/T region in a manner similar to what is seen for application of hydrostatic pressureto superconducting samples of Ca(Fe − x Co x ) As . Given that previous work showed thatboth cT and AFM/ORTH phases are much more sensitive to changes in the c -axis than tochanges in the ab -axis, this can be understood based on the fact that Rh-substitution sup-presses c -lattice parameter more rapidly than either Co- or Ni-substitution. The cT phaseline starts near 70 K at x = 0.028 and reaches 140 K at x = 0.065, where the transitionbecomes broadened as also seen for the T A/Q = 960 ◦ C, high substitution levels. The threelow temperature states can be seen in the phase diagram presented in Fig. 9d. Note thatat x = 0.02, low field magnetic susceptibility shows superconducting signal with screeningof more than 60% of 1/4 π , whereas resistance data (which was taken on the same pieceof sample) shows upward turning upon cooling indicating AFM/ORTH transition. Giventhat x = 0.02 is at the phase boundary, it is very likely that part of the sample transformsinto SC/PM/T phase and the other part of the sample transforms into AFM/ORTH phase.The other possibility is the coexistence of superconductivity and antiferromagnetism. Thisscenario is unlikely because the AFM/ORTH phase transition in this compound remainsquite first order even though it is suppressed to around 50 K, as indicated by the sharpnessof the resistive signature of the transition. Therefore, there would not be enough magneticfluctuations, which are vital for the emergence of the unconventional superconductivity inthe iron pnictides according to the current theories, as well as our earlier experimentson Ca(Fe − x Co x ) As , to support superconductivity in the AFM/ORTH state.Figure 10 presents T - x phase diagrams for different annealing/quenching temperaturesand T - T A/Q phase diagrams for different Rh concentrations. Similar to what we did forNi-substitution, we assembled these data and constructed a 2D phase diagram for the15 M / H ( / ) T (K) SC/PM/T cTAFM/ORTH (d) T S / T N - M T S / T N - R T C - M T cT - M T r an s i t i on T ( K ) x N/PM/T R / R ( K ) T (K) Rh - T A/Q = 400(cid:176)C M H ( - e m u / m o l e ) T (K) (a) x = FIG. 9: (Color online) Temperature dependent (a) magnetic susceptibility with field applied parallelto the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01 Tapplied perpendicular to the c axis, (c) normalized electrical resistance, and (d) phase diagramof transition temperature T vs Rh concentration x of Ca(Fe − x Rh x ) As samples with T A/Q =400 ◦ C. Susceptibility data in (a) have been offset from each other by an integer multiple of 3 × − emu/mole for clarity. base-temperature states of the Ca(Fe − x Rh x ) As system, as presented in Fig. 8. The T A/Q - x , 2D phase diagram of Rh-substitution is significantly different from that of Co-substitution. This is in contrast to the case of Ba(Fe − x T x ) As , where the phase diagramsfor Co- and Rh-substitutions are almost identical. In the case of Ca(Fe − x Rh x ) As theAFM/ORTH phase is suppressed faster than it is for Co-substitution and the cT phaseis much more pervasive in case of Rh-substitution, appearing for all annealing/quenchingtemperatures for substitutions level above 3%. Both of these changes can be understood16ased on the fact that Rh-substitution suppresses the c -lattice parameter more rapidly thanCo (or Ni) substitution. A consequence of the enhanced stabilization of the cT phase forlow T A/Q values is (i) the complete absence on the N/PM/T phase and (ii) the SC/PM/Tregion for Rh-substitution is substantially shrunk, or truncated, compared with that forCo-substitution. Given that (i) current theories and experiments indicate that the spinfluctuations play an important role for the appearance of unconventional superconductivityin the iron pnictides; and (ii) spin fluctuations are completely suppressed in the cT phasein CaFe As , it is likely that the superconductivity is limited by the pervasive cTphase in the Ca(Fe − x Rh x ) As system. Critical c -lattice parameter The cT phase transition is driven by an increasing overlap of interlayer As orbitals. While it was suggested that As-As interlayer separation appears to be the key parametercontrolling the volume collapse when comparing members of the ThCr Si structure, it is conceivable that, for substitutions to CaFe As , there might also be a critical roomtemperature c -lattice parameter value. Given that As-As interlayer separation is hard tomeasure, a critical room temperature c -lattice parameter value can give an easy evaluationof whether the system will transform into the cT phase or not. In order to assess the extentto which such a critical value can be inferred, we plotted the c -lattice parameter versussubstitution level, x , for all three substitutions with various T A/Q , as shown in Fig. 11. Rareearth substitution data from literature, as well as data for Sn-grown, x = 0 CaFe As underpressure, are also presented for comparison. It can be seen that the room temperature c -lattice parameter can be divided into three regions: (i) below 11.64 ˚ A , where all the samplestransform into cT phase at low temperature; (ii) above 11.73 ˚ A , where all the samples havenon-cT phase as low temperature ground state; (iii) between 11.64 ˚ A and 11.73 ˚ A , wheredetails, such as temperature dependence of thermal contraction, amount of internal strain,specific type of substitution, etc., become important for determining the low temperaturestructural state. Note that all the rare earth substituted samples fall into the last categorywhich is consistent with the fact that detailed As-As interlayer separation determines theground state. The As-As interlayer separation of the Ce-substituted samples with x = 0.16,when extrapolated to base temperature assuming a constant temperature dependence, is17
00 600 800 1000050100150
SC/PM/T cTN/PM/T (d)
Rh - x = 0.023 T r an s i t i on T ( K ) T A/Q ((cid:176)C)
N/PM/T cT (b)
Rh - T A/Q = 960(cid:176)C T r an s i t i on T ( K ) x N/PM/T cTAFM/ORTH (a)
Rh - T A/Q = 500(cid:176)C T r an s i t i on T ( K ) x T S / T N - M T S / T N - R T C - M T cT - M
400 600 800 1000050100150 cTN/PM/TAFM/ORTH (c)
Rh - x = 0.015 T r an s i t i on T ( K ) T A/Q ((cid:176)C)
FIG. 10: (Color online) phase diagram of (a) transition temperature T vs Rh concentration x ofCa(Fe − x Rh x ) As samples with T A/Q = 500 ◦ C, (b) transition temperature T vs Rh concentra-tion x of Ca(Fe − x Rh x ) As samples with T A/Q = 960 ◦ C, (c) transition temperature T vs anneal-ing/quenching temperature T A/Q of Ca(Fe − x Rh x ) As samples with Rh concentration x = 0.015,and (d) transition temperature T vs annealing/quenching temperature T A/Q of Ca(Fe − x Rh x ) As samples with Rh concentration x = 0.023. just above the claimed critical value. On the other hand, the room temperature c -latticeparameter, 11.65 ˚ A , is also on the edge of the last region, showing good agreement with thecriteria of As-As interlayer separation. The data for Sn-grown, x = 0 sample also fit ourcriteria very well. Under the ambient pressure, the room temperature c -lattice parameterfalls into the second category with the low temperature state being a AFM/ORTH phase,whereas under the pressure of 0.62 GPa, the c -lattice parameter, when extrapolated to roomtemperature, falls into the first category with the low temperature state being a cT phase.18 .00 0.05 0.10 0.15 0.2011.411.511.611.711.8 cT Ni Rh Co Re P = 0.62 GPa non-cT Ni Rh Co Re P = 0 GPa c ( ¯ ) x FIG. 11: (Color online) c -lattice parameter versus substitution level of all three substitutions. Dataof rare earth substitution from Ref. and data for Sn-grown pure CaFe As under pressure fromRef. are also included for comparison. Annealing time dependence
For earlier work on both pure CaFe As and Ca(Fe − x Co x ) As , we performed sys-tematic studies of effects of annealing time for various T A/Q and showed that the effects ofannealing were established rather quickly (t <
24 h) for T A/Q of interest. In addition, forboth pure CaFe As and Ca(Fe − x Co x ) As we found that longer annealing time did notsignificantly change the T - T A/Q phase diagrams indicating that there was only one salientannealing process with a single characteristic time. As an example, virtually identical phasediagrams of Ca(Fe − x Co x ) As for T A/Q = 500 ◦ , assembled from two different sets of data,1-day anneal and 7-day anneal are presented in Fig. 12.Ni- and Rh-substitutions appear to be different. Although 1-day annealing gives familiarphase diagrams, they change with longer annealing times. Figure 13 presents the phasediagrams for Ni-substitution for T A/Q = 500 ◦ C with different annealing time sequences.As can be seen, for samples annealed for seven days, the AFM/ORTH phase transitionis suppressed more slowly and the SC/PM/T phase is only stabilized for a slightly higherNi concentration level. However, the reproducibility with respect to annealing/quenching19 .00 0.02 0.04050100150
Co - T A/Q = 500(cid:176)C T S /T N c S /T N c T r an s i t i on T ( K ) x FIG. 12: (Color online) Phase diagrams of transition temperature T vs Co concentration x as-sembled from magnetic susceptibility data, for Ca(Fe − x Co x ) As samples with T A/Q = 500 ◦ C. Filled symbols are inferred from data from samples with 1-day annealing and open symbols areinferred from data from samples with 7-day annealing. history seems to be preserved. We took these 7-day, 500 ◦ C annealed samples, resealed them,annealed/quenched at 800 ◦ C trying to bring the samples back to a state that is close to T A/Q = 960 ◦ C samples, and then annealed again at 500 ◦ C for one day and quenched. Afterthis series of annealing, the T - x phase diagram is similar to that is seen for the initial 1-dayannealing, indicating that whatever process is taking place over this longer time scale, it isreversible. These data imply that (i) there is more than one salient annealing time, but that(ii) there is clear reversibility and reproducibility.Even larger effects of a longer annealing time are observed for Rh-substitution as shownin Fig. 14. It can be seen that the AFM/ORTH phase transition is initially suppressedmore slowly for the 7-day annealed/quenched samples than for the 1-day annealed/quenchedsamples. In addition, the SC/PM/T ground state is stabilized at low temperature for the 7-day annealed/quenched samples with substitution level of 3.7% and higher. This is in starkcontrast to what has been seen for 1-day annealed/quenched Ca(Fe − x Rh x ) As compounds,where the cT phase is found for high substitution level and no superconductivity is revealed.Again we resealed these 7-day annealed samples, annealed/quenched at 800 ◦ C, and then20
Ni - T
A/Q = 500(cid:176)C T r an s i t i on T ( K ) T S / T N T C x FIG. 13: (Color online) Phase diagrams of transition temperature T vs Ni concentration x assem-bled from magnetic susceptibility data, for Ca(Fe − x Ni x ) As samples with T A/Q = 500 ◦ C. (a)1-day annealing, (b) 7-day annealing, (c) 1-day annealing after a series of annealing described inthe text. For comparison, data in (a) are repeated in (c) with open symbols. annealed at 500 ◦ C for one day and quenched. As seen for Ni-substituted samples, after thisseries of annealing, the initial “1-day anneal” phase diagram is recovered, illustrating clearreversibility and reproducibility.The clear difference between effects of 1-day and 7-day annealing, as well as the clearreversibility and reproducibility, can also be seen in the c -lattice parameter data from the21 T cT Rh - T
A/Q = 500(cid:176)C S /T N T C (b)7-day annealing T r an s i t i on T ( K ) (c) x FIG. 14: (Color online) Phase diagrams of transition temperature T vs Rh concentration x assem-bled from magnetic susceptibility data, for Ca(Fe − x Rh x ) As samples with T A/Q = 500 ◦ C. (a)1-day annealing, (b) 7-day annealing, (c) 1-day annealing after a series of annealing described inthe text. For comparison, data in (a) are repeated in (c) with open symbols.
Ca(Fe − x Rh x ) As system as presented in Fig. 15. The c -lattice parameter is suppressed byRh-substitution much less rapidly for 7-day, 500 ◦ C annealed/quenched samples, than for 1-day, 500 ◦ C annealed/quenched samples. After a series of further thermal treatment, we couldbring it back to the behavior similar to what is seen for 1-day, 500 ◦ C annealed/quenchedsamples.The origin of this annealing time dependence of the physical properties is still unknown.22 .00 0.02 0.04 0.0611.6011.6511.7011.75
Rh - T
A/Q = 500(cid:176)C c ( ¯ ) x FIG. 15: (Color online) Room temperature c -lattice parameter of Ca(Fe − x Rh x ) As samples with T A/Q = 500 ◦ C as a function of measured Rh concentration, x , for 1-day annealing (red), 7-dayannealing (black) and 1-day annealing after a series of annealing described in the text (blue). One possibility is that there are two salient time scales. One timescale for the small, excessof FeAs associated with the CaFe As width of formation to go in and out of the CaFe As matrix, as we proposed based on our T - T A/Q phase diagram and TEM results.
Anothertimescale for some Fe/Ni (and Fe/Rh) segregation. Note that this is only speculation butwould fit the data. As we change annealing times around the second time scale, we wouldchange the Rh/Ni (or RhAs/NiAs) content and therefore change the phase diagram as well asthe c -lattice parameter in a reversible manner. The fact that Co-substitution does not showthe same annealing time dependence raises the question of what the differences betweensolubility of Co and Rh/Ni or CoAs and RhAs/NiAs in the CaFe As matrix are. Moredetailed microscopic study, such as high resolution TEM, will be needed to provide furtherinsight into this issue. IV. SUMMARY
We report systematic studies of the combined effects of annealing/quenching tempera-ture and Ni/Rh-substitution on the physical properties of CaFe As . We constructed two-23imensional phase diagrams for the low-temperature states for both systems to map out therelations between possible ground states and then compared with that of Co-substitution.Ni-substitution, which brings one more extra electron per substituted atom and suppressesthe c -lattice parameter at roughly the same rate as Co-substitution, leads to similar changesin the Ca(Fe − x Ni x ) As phase diagram as were seen when comparing the Ba(Fe − x Co x ) As and Ba(Fe − x Ni x ) As phase diagrams: similar suppression of the AFM/ORTH phase but amore rapid suppression of the SC/PM/T phase for Ni-substitution. On the other hand, Rh-substitution, which brings the same amount of extra electrons but suppresses the c -latticeparameter more rapidly that Co-substitution, has a very different phase diagram from thatof Ca(Fe − x Co x ) As : Rh-substitution suppresses the AFM/ORTH phase more rapidly thanCo-substitution, but more dramatically, the cT phase is stabilized over a much greater regionof the x - T A/Q phase space, truncating the SC/PM/T region. In addition to the differencesin phase diagrams, we also found different behavior in both systems related to annealingtime compared to Co-substitution. We propose that for Ni- and Rh-substitution, there is asecond, reversible process taking place on a longer time scale, but at the current time we donot know its microscopic origin.
Acknowledgments
S.Ran acknowledges Anton Jesche for help on the x-ray measurement. Authors acknowl-edge Matthew Kramer and Lin Zhou for usefully discussions. Work at the Ames Laboratorywas supported by the Department of Energy, Basic Energy Sciences, Division of MaterialsSciences and Engineering under Contract No. DE-AC02-07CH11358. S.L.B. acknowledgespartial support from the State of Iowa through Iowa State University.24 ppendix
This appendix presents the magnetic susceptibility and resistance data for both Ni- andRh-substitutions that were used to construct phase diagrams presented and discussed in themain text. For as grown samples (which were quenched from 960 ◦ C), due to the violentstructure phase transition associated with the cT phase transition, the resistance measure-ments suffer from cracking and contact problems. Therefore only magnetic susceptibilitydata are presented.
Ni-substitution T A/Q = 500 ◦ C Figure 16 presents the data used to construct the T - x phase diagram for Ni-substitutionwith T A/Q = 500 ◦ C shown in Fig. 7a. The AFM/ORTH transition is suppressed com-pletely between x = 0.017 and 0.019. Sample with x = 0.019 shows significant amount ofdiamagnetism with T c around 6 K. Ni-substitution T A/Q = 960 ◦ C Figure 17 presents the data used to construct the T - x phase diagram for Ni-substitutionwith T A/Q = 960 ◦ C shown in Fig. 7b. The drop in susceptibility is suppressed to lowertemperature as Ni substitution level is increased.
Ni-substitution x = 0.021 Figure 18 presents the data used to construct the T - T A/Q phase diagram for Ni-substitution with x = 0.021 shown in Fig. 7c. The AFM/ORTH phase transition takesplace for T A/Q ◦ C and the cT phase is stabilized for T A/Q > ◦ C. No bulk super-conductivity is observed for any T A/Q . Ni-substitution x = 0.026 Figure 19 presents the data used to construct the T - T A/Q phase diagram for Ni-substitution with x = 0.026 shown in Fig. 7c. Sample with T A/Q = 400 ◦ C shows su-25
100 200 3005101520 H = 1T ll c
Ni - T A/Q = 500(cid:176)C M H ( - e m u / m o l e ) T (K) x = 0 5 10 15-1.0-0.8-0.6-0.4-0.20.0 H = 0.01T cZFC M / H ( / ) T (K) R / R ( K ) T (K) FIG. 16: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Ni x ) As samples with T A/Q = 500 ◦ C. Susceptibility data in (a) have been offset from each other by aninteger multiple of 1 × − emu/mole for clarity. perconductivity with screening of 80% of 1/4 π . The cT phase is stabilized for T A/Q > ◦ C. Rh-substitution T A/Q = 500 ◦ C Figure 20 presents the data used to construct the T - x phase diagram for Rh-substitutionwith T A/Q = 500 ◦ C shown in Fig. 10a. The AFM/ORTH phase transition is suppressed to95 K by x = 0.011 and the cT phase is stabilized by x = 0.015. No bulk superconductivity26
100 200 300051015202530
Ni - T
A/Q = 960(cid:176)C
H = 1T ll c M H ( - e m u / m o l e ) T (K) x = FIG. 17: (Color online) Temperature dependent magnetic susceptibility with field applied parallelto the c axis of Ca(Fe − x Ni x ) As samples with T A/Q = 960 ◦ C. Data have been offset from eachother by an integer multiple of 2 × − emu/mole for clarity. is observed for any Rh substitution level. Rh-substitution T A/Q = 960 ◦ C Figure 21 presents the data used to construct the T - x phase diagram for Rh-substitutionwith T A/Q = 960 ◦ C shown in Fig. 10b. With increasing Rh substitution level, the transitiontemperature of the cT phase is enhanced significantly and the feature associated with thephase transition becomes significantly broadened.
Rh-substitution x = 0.015 Figure 22 presents the data used to construct the T - T A/Q phase diagram for Rh-substitution with x = 0.015 shown in Fig. 10c. The AFM/ORTH phase transition takesplace for T A/Q < ◦ C and the cT phase is stabilized for T A/Q > ◦ C. No bulk super-conductivity is observed for any T A/Q . 27 M / H ( / ) T (K) M / H ( / ) T (K) R / R ( K ) T (K) M H ( - e m u / m o l e ) T (K)Ni - x = 0.021 FIG. 18: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Ni x ) As samples with Ni concentration x = 0.021. Susceptibility data in (a) have been offset from eachother by an integer multiple of 5 × − emu/mole for clarity. Rh-substitution x = 0.023 Figure 23 presents the data used to construct the T - T A/Q phase diagram for Rh-substitution with x = 0.023 shown in Fig. 10d. Superconductivity with full screening isobserved for samples with T A/Q = 350 ◦ C and 400 ◦ C. The cT phase is stabilized for T A/Q > ◦ C. 28
100 200 3000102030 H = 1T ll c Ni - x = 0.026 M H ( - e m u / m o l e ) T (K) R / R ( K ) T (K) M / H ( / ) T (K) FIG. 19: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Ni x ) As samples with Ni concentration x = 0.026. Susceptibility data in (a) have been offset from eachother by an integer multiple of 5 × − emu/mole for clarity. M / H ( / ) T (K) M / H ( / ) T (K)
Rh - T A/Q = 500(cid:176)C
H = 1T ll c M H ( - e m u / m o l e ) T (K) x = R / R ( K ) T (K) FIG. 20: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Rh x ) As samples with T A/Q = 500 ◦ C. Susceptibility data in (a) have been offset from each other by aninteger multiple of 3 × − emu/mole for clarity.
100 200 300051015202530
Rh - T A/Q = 960(cid:176)C
H = 1T ll c M H ( - e m u / m o l e ) T (K) x = FIG. 21: (Color online) Temperature dependent magnetic susceptibility with field applied parallelto the c axis of Ca(Fe − x Rh x ) As samples with T A/Q = 960 ◦ C . Data have been offset from eachother by an integer multiple of 3 × − emu/mole for clarity.
100 200 3000510152025303540 H = 1T ll c
Rh - x = 0.015 M H ( - e m u / m o l e ) T (K) R / R ( K ) T (K) M / H ( / ) T (K) M / H ( / ) T (K)
FIG. 22: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Ni x ) As samples with Rh concentration x = 0.015. Susceptibility data in (a) have been offset from eachother by an integer multiple of 4 × − emu/mole for clarity. M / H ( / ) T (K) Rh - x = 0.023 M H ( - e m u / m o l e ) T (K) R / R ( K ) T (K) FIG. 23: (Color online) Temperature dependent (a) magnetic susceptibility with field appliedparallel to the c axis, (b) low-field magnetic susceptibility measured upon ZFC with a field of 0.01T applied perpendicular to the c axis, and (c) normalized electrical resistance of Ca(Fe − x Ni x ) As samples with Rh concentration x = 0.023. Susceptibility data in (a) have been offset from eachother by an integer multiple of 4 × − emu/mole for clarity. N. Ni, S. Nandi, A. Kreyssig, A. I. Goldman, E. D. Mun, S. L. Bud’ko, and P. C. Canfield,Phys. Rev. B , 014523 (2008). A. I. Goldman, D. N. Argyriou, B. Ouladdiaf, T. Chatterji, A. Kreyssig, S. Nandi, N. Ni, S. L.Bud’ko, P. C. Canfield, and R. J. McQueeney, Phys. Rev. B , 100506(R) (2008). A. Kreyssig, M. A. Green, Y. Lee, G. D. Samolyuk, P. Zajdel, J. W. Lynn, S. L. Bud’ko, M. S.Torikachvili, N. Ni, S. Nandi, J. B. Leao, S. J. Poulton, D. N. Argyriou, B. N. Harmon, R. J.McQueeney, P. C. Canfield, and A. I. Goldman, Phys. Rev. B , 184517 (2008). A. I. Goldman, A. Kreyssig, K. Prokes, D. K. Pratt, D. N. Argyriou, J. W. Lynn, S. Nandi, S.A. J. Kimber, Y. Chen, Y. B. Lee, G. Samolyuk, J. B. Leao, S. J. Poulton, S. L. Bud’ko, N. Ni,P. C. Canfield, B. N. Harmon, and R. J. McQueeney, Phys. Rev. B , 024513 (2009). M. S. Torikachvili, S. L. Bud’ko, N. Ni, and P. C. Canfield, Phys. Rev. Lett. , 057006 (2008). W. Yu, A. A. Aczel, T. J. Williams, S. L. Bud’ko, N. Ni, P. C. Canfield, and G. M. Luke, Phys.Rev. B , 020511 (2009). P. C. Canfield, S. L. Bud’ko, N. Ni, A. Kreyssig, A. I. Goldman, R. J. McQueeney, M. S.Torikachvili, D. N. Argyriou, G. Luke, and W. Yu, Physica C , 404 (2009). H. Lee, E. Park, T. Park, V. A. Sidorov, F. Ronning, and E. D. Bauer, and J. D. Thompson,Phys. Rev. B , 024519 (2009). K. Prokes, A. Kreyssig, B. Ouladdiaf, D. K. Pratt, N. Ni, S. L. Bud’ko, P.C.Canfield,R.J.McQueeney, D.N.Argyriou, and A. I. Goldman, Phys. Rev. B , 180506 (2010). T. Park, E. Park, H. Lee, T. Klimczuk, E. D. Bauer, F. Ronning, and J. D. Thompson, J. Phys.Condens. Matter , 322204 (2008). S. Ran, S. L. Bud’ko, D. K. Pratt, A. Kreyssig, M. G. Kim, M. J. Kramer, D. H. Ryan, W. N.Rowan-Weetaluktuk, Y. Furukawa, B. Roy, A. I. Goldman, and P. C. Canfield, Phys. Rev. B , 144517 (2011). S. Ran, S. L. Bud’ko, W. E. Straszheim, J. Soh, M. G. Kim, A. Kreyssig, A. I. Goldman, andP. C. Canfield, Phys. Rev. B , 224528 (2012). R. S. Dhaka, R. Jiang, S. Ran, S. L. Bud’ko, P. C. Canfield, M. Tomic, R. Valenti, Y. Lee, B.N. Harmon, A. Kaminski, Phys. Rev. B , 020511(R) (2014). S. L. Bud’ko, S. Ran, and P. C. Canfield, Phys. Rev. B , 064513 (2013). E. Gati, S. K¨ohler, D. Guterding, B. Wolf, S. Kn¨oner, S. Ran, S. L. Bud’ko, P. C. Canfield, andM. Lang , Phys. Rev. B , 220511(R) (2012). N. Ni, A. Thaler, A. Kracher, J. Q. Yan, S. L. Bud’ko, and P. C. Canfield, Phys. Rev. B ,024511 (2009). P. C. Canfield, S. L. Bud’ko, N. Ni, J. Q. Yan, and A. Kracher, Phys. Rev. B , 060501(R)(2009). N. Ni, A. Thaler, J. Q. Yan, A. Kracher, E. Colombier, S. L. Bud’ko, P. C. Canfield, and S. T.Hannahs, Phys. Rev. B , 024519 (2010). P. C. Canfield and S. L. Bud’ko, Annu. Rev. Condens. Matter Phys. , 27 (2010). Z. Fisk, J.P. Remeika, in: K.A. Gschneider, L. Eyring (Eds.), Handbook on the Physics andChemistry of Rare Earths, Vol. , Elsevier, Amsterdam, 1989. P. C. Canfield and Z. Fisk, Philos. Mag. B , 1117 (1992). P. C. Canfield, Solution Growth of Intermetallic Single Crystals: A Beginners Guide, Propertiesand Applications of Complex Intermetalltics (World Scientific Publishing Co. Pte. Ltd., 2010),pp. 93 —111. A. S. Sefat, R. Jin, M. A. McGuire, B. C. Sales, D. J. Singh, and D. Mandrus, Phys. Rev. Lett. R. Hu, S. Ran, S. L. Bud’ko, W. E. Straszheim, P. C. Canfield, Philos. Mag , 3113 (2012). N. Ni, M. E. Tillman, J.-Q. Yan, A. Kracher, S. T. Hannahs, S. L. Bud’ko, and P. C. Canfield,Phys. Rev. B , 214515 (2008). S. R. Saha, N. P. Butch, T. Drye, J. Magill, S. Ziemak, K. Kirshenbaum, P. Y. Zavalij, J. W.Lynn and J. Paglione, Phys. Rev. B , 024525 (2012). I. I. Mazin, D. J. Singh, M. D. Johannes, and M. H. Du, Phys. Rev. Lett. , 057003 (2008). R. M. Fernandes and J. Schmalian, Phys. Rev. B , 014521 (2010). T. Yildirim, Phys. Rev. Lett. D. K. Pratt, Y. Zhao, S. A. J. Kimber, A. Hiess, D. N. Argyriou, C. Broholm, A. Kreyssig, S.Nandi, S. L. Bud’ko, N. Ni, P. C. Canfield, R. J. McQueeney and A. I. Goldman, Phys. Rev. B , 060510(R) (2009). J. H. Soh, G. S. Tucker, D. K. Pratt, D. L. Abernathy, M. B. Stone, S. Ran, S. L. Bud’ko, P.C. Canfield, A. Kreyssig, R. J. McQueeney, and A. I. Goldman, Phys. Rev. Lett. Y. Furukawa, B. Roy, S. Ran, S. L. Bud’ko, P. C. Canfield, Phys. Rev. B , 121109(R) (2014). R. Hoffmann and C. Zheng, J. Phys. Chem. , 4175 (1985)., 4175 (1985).