The effects of electron correlation and spin-orbit coupling in the isovalent Pd-doped superconductor SrPt P 3 Kangkang Hu,1,2 Bo Gao,1 Qiucheng Ji,1 Yonghui Ma,1,3 Wei Li,1,4,∗ Xuguang Xu,3 Hui Zhang,5 Gang Mu,1,† Fuqiang Huang,5 Chuanbing Cai,2 Xiaoming Xie,1 and Mianheng Jiang1,3 1State Key Laboratory of Functional Materials for Informatics and Shanghai Center for Superconductivity, Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences, Shanghai 200050, China 2Shanghai Key Laboratory of High Temperature Superconductors, Shanghai University, Shanghai 200444, China 3School of Physical Science and Technology, ShanghaiTech University, Shanghai 201210, China 6 4CAS-Shanghai Science Research Center, Shanghai 201203, China 1 5CAS Key Laboratory of Materials for Energy Conversion, 0 Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai 200050, China 2 (Dated: January 13, 2016) n We present a systematical study on the roles of electron correlation and spin-orbit coupling in a theisovalentPd-dopedsuperconductorSrPt P.Byusingsolidstatereactionmethod,wefabricated 3 J thestrongspin-orbitcouplingsuperconductorsSr(Pt Pd ) PwithstrongelectroncorrelatedPd 1−x x 3 2 dopant of the 4d orbital. As increasing the isovalent Pd concentrations without introducing any 1 extraelectron/holecarriers,thesuperconductingtransitiontemperatureT decreasesmonotonously, c whichsuggeststheexistenceofcompetitionbetweenspin-orbitcouplingandelectroncorrelationin ] thesuperconductingstate. Inaddition,theelectronicbandstructurecalculationsdemonstratethat n the strength of electron susceptibility is suppressed gradually by the Pd dopant suggesting the o incompatible relation between spin-orbit coupling and electron correlation, which is also consistent c - with experimental measurements. Our results provide significant insights in the natures of the r interplay between the spin-orbit coupling and the electron correlation in superconductivity, and p may pave a way for understanding the mechanism of superconductivity in this 5d-metal-based u compound. s . t a PACSnumbers: 74.70.Wz,75.47.-m,71.70.Di,74.25.Jb m - d I. INTRODUCTION basedsuperconductorsSrPt3P,wherethestrengthofthe n spin-orbitcouplingandelectroncorrelationcanbetuned o by the Pd concentration without introducing any extra c The study of strong interplay among charge, spin, or- electron or hole carriers into the system. This physics [ bital, and lattice degrees of freedom in transition metal is quite different from the case that of the Lanthanum- 1 compounds has triggered enormous research interests in based superconductor LaPt3P, where the extra electron v the communities of condensed matter physics and mate- is injected into the system as strontium is replaced by 2 rial physics. One of the most prominent example is the Lanthanum leading to the decrease of the total density 8 unconventionalhigh-transitiontemperature(high-T )su- of states (DOS) at the Fermi level, which results in the 7 c perconductivity induced by the strong electron correla- suppression of superconducting transition temperature 2 0 tion in the copper oxide and iron-based superconduc- Tc [7]. . tors[1,2]. Inthosematerials,apparently,thespindegree 1 In this paper, the Pd dopant with strong electron offreedomplaysavitalruleandtheorbitaldegreeoffree- 0 correlation of 4d orbital was successfully substituted to 6 dom is decoupled from that of spin. However, in strong the site of Pt in strong spin-orbit coupling superconduc- 1 spin-orbit coupling superconductors, such as the recent : discoveredplatinum-basedsuperconductorsAPt P(A= tor SrPt3P, which was confirmed by the x-ray diffrac- v 3 tion measurements. The crystal lattice was found to Ca, Sr and La) [3–10], the systems display weak electron i X correlation effect. The relation between strong electron shrink along the c-axis and expand along the a-axis monotonously as increasing the Pd concentrations. Im- r correlation and spin-orbit coupling still remains unclear. a portantly, we find that the superconducting transition To clarify the nature of the interplay between spin-orbit temperature T decreases with Pd doping, which could couplingandelectroncorrelationinsuperconductorsisa c not be attributed to the physics of changes of the DOS crucial issue not only in condensed matter physics, but at the Fermi level and the impurity scattering. Thus, it alsoinmaterialscience. Motivatedbythisissue,wefocus suggests that the system exhibits a competition between ourgreatattentionsontheisovalentPddopedplatinum- electron correlation and spin-orbit coupling in our sys- tem. Suchacompetingrelationisalsoconsistentwiththe first-principles calculation, which show that the strength ∗[email protected] of electron susceptibility is suppressed gradually as in- †[email protected] creasing Pd dopants. 2 II. EXPERIMENTS Thesamplesinthisworkwerepreparedviasolidstate reaction from the pure elements. [11] Firstly, we put sto- ichiometricamountsofplatinumpowder(purity99.97%, Alfa Aesar), red phosphorus powder (purity 99.9%, Al- addin), and strontium pieces (purity 98%+, Alfa Aesar) together and ground them in a mortar. After that, the mixture was pressed into a small pellet and then sealed in a clean vacuum quartz tube. All the weighing and mixingprocedureswerecarriedoutinagloveboxwitha protectiveargonatmosphere. Thetubewasheatedupto 400 ◦C and held for 10 hours to prevent red phosphorus from volatilizing so quickly, and calcined at 900 ◦C for 2 days. The sintered pellet was reground and further an- nealedat900◦Cwithinanargon-filledquartztubesfor3 days. ThedopedsamplesSr(Pt Pd ) Pwereprepared 1−x x 3 with adding corresponding amount of palladium powder (purity 99.95%, Alfa Aesar) using the same method as FIG. 1: (color online) (a) X-ray diffraction patterns for the mentioned above. Sr(Pt Pd ) P samples with 0 ≤ x ≤ 0.4. Small amount 1−x x 3 The structure of the obtained samples were checked of impurity phases are indexed by the blue asterisks. (b) An using a DX-2700 type powder x-ray diffractometer. The enlarged view of the (220) peak. It is clear that the peak magnetic susceptibility measurements were carried out shifts monotonously to the left direction with doping. (c) on the magnetic property measurement system (Quan- CrystalstructureofSrPt3P.Pdelementwassubstitutedinto the site of Pt in this work. (d) Doping dependence of the tum Design, MPMS 3). The electrical resistance was lattice constants along a-axis and c-axis. measured using a four-probe technique on the phys- ical property measurement system (Quantum Design, PPMS). a-axis with the increase of doping. This tendency is very similar to that observed in 4d- and 5d-metal-doped ironarsenidesSrFe M As (M=Rh,Ir,Pd),[13]which III. RESULTS AND DISCUSSIONS 2−x x 2 proves that Pd atoms take the place of Pt atoms leading to the alteration of cell parameters. In the high doping A. Crystal Structure region above 0.4, the variation of the lattice parameters becomes gentle, indicating the limit of the chemical sub- The crystalstructure of this systemhas been reported stitution. We note that this is a common phenomenon by T. Takayama et al. previously [3]. Here we concen- in some chemical doped materials and such a saturated trate to the influences on the crystal lattice by the Pd features has been reported elsewhere. [14] substitution. Forthisreason,wemeasuredthepowderX- ray diffraction (XRD) on this series of materials and the diffraction patterns are shown in Fig. 1 (a). It is found that the main diffraction peaks can be indexed to the B. Superconducting Properties tetragonal structure with the space group P4/nmm as shown in Fig. 1(c). The black line represents the parent In order to study the effect of Pd-doping on the su- compound SrPt P. The asterisks refer to some unknown perconductingproperties,weperformedthetemperature 3 impurities which was also observed by T. Takayama et dependent magnetization and resistivity measurements al. [3] With the increase of the amount of Pd (x), there on all the samples we synthesized. The temperature de- is no obvious increase of the impurity phases. However, pendence of dc magnetic susceptibility is shown in Fig. the positions of diffraction peaks start to move appar- 2(a), where we can see a sharp decline to the negative ently when x changes from 0 to 0.4, which points to the sides of the data for each sample with different doping. variation of the size of the crystal lattice. This tendency The clear diamagnetic signal indicates the occurrence of can be seen clearly in Fig. 1(b), where we enlarge the the superconducting transition and the onset transition region near the (220) diffraction peak as an example. pointdefinesthecriticaltransitiontemperatureT . With c To investigate the influences on the crystal lattice by theincreaseofamountofPd, thevalueofT reducesap- c the Pd substitution quantitatively, we obtained the lat- parently. This tendency is confirmed by the resistivity tice parameters by fitting the XRD data using the soft- data, which is shown in Fig. 2(b). Clear superconduct- ware Powder-X. [12] The results are shown in Fig. 1(d). ing transitions to zero resistance were observed on all As is shown in the graph, the crystal lattice shows a thesampleswithdifferentdopinglevels. Theonsettran- shrinkage along the c-axis, while it expands along the sition temperatures determined from this figure, along 3 suppression of T [17]. c Since the experimental observation of the suppression of the T does not originate from the changes of DOS at c the Fermi level and the impurity scattering effect, the T suppression is likely to be related to the interplay c between the strong spin-orbit coupling and the electron correlation. As we have known, the spin-orbit coupling strength is proportional to Z4 (where Z is the atomic number; Z =78 for Pt and Z =46 for Pd) [18], the Pt Pd ratio γ = (ZPt)4 = 8.3 and consequently the spin-orbit coupling strZePndgth will be decreased dramatically when the Platinum is substituted by Palladium. In addition, the bandwidth of Pd 4d orbitals is narrower than that of Pt 5d orbitals making the electron correlation in Pd 4d orbital electrons is stronger than that in Pt 5d or- bital electrons. Therefore, when the enhancement of the strength of electron correlation by Palladium dopants is FIG. 2: (color online) (a) Temperature dependence of the comparable to that of spin-orbit coupling, the interplay dc magnetic susceptibility for Sr(Pt Pd ) P samples with 1−x x 3 between electron correlation and the spin-orbit coupling 0 ≤ x ≤ 0.4. The data were measured via the zero-field- will play a crucial rule in superconductivity and affect cooling(ZFC)modeunderthefieldof10Oe. (b)Temperature the superconducting transition temperature T . dependence of the normalized resistivity. (c) Doping depen- c dence of the onset superconducting transition temperature. The blue and purple lines were defined by the magnetization measurements and the resistivity measurements respectively. C. First-Principles Calculations In order to clarify the interplay between electron with that determined from the M −T curves, are dis- correlation and spin-orbit coupling and the origins for played in Fig. 2(c). The blue symbols were obtained the suppression of T by Pd doping in superconduc- c from the magnetization measurements while the purple tor Sr(Pt Pd ) P theoretically, we carried out the 1−x x 3 onesweredefinedbytheresistivitymeasurements. Itcan first-principles band structure calculations on the Pd be seen that the two sets of Tc values evolute parallelly dopant dependent samples with x = 0, x = 0.17, and and decrease linearly with doping. x = 0.33. The calculations were performed by using Generally speaking, the suppression of superconduct- the pseudopotential-based code VASP [19] within the ing transition temperature T is usually related to the Perdew-Burke-Ernzerhof [20] generalized gradient ap- c changes of DOS at Fermi level, as has been reported in proximation. Throughout the theoretical calculations, theLanthanumreplacedsuperconductorLaPt Pbycom- a 500 eV cutoff in the plane wave expansion and a 3 paringwithSrPt P[7]. However,thecalculatedDOSby 12×12×12 Monkhorst(cid:126)k grid were chosen to ensure the 3 meansofdensityfunctionaltheorydemonstratedthatthe calculation with an accuracy of 10−5 eV. The atomic co- DOS around the Fermi level remains almost the same ordinates were obtained by the relaxation based on the fordifferentconcentrationsofPddopants[seeFig.3(d)], lattice parameters from experiments. Additionally, the whichwillbediscussedinthenextsectionindetail. This spin-orbitcouplinghadbeenincludedthroughoutthecal- result is reasonable since the Palladium is isovalent to culations. Platinum making the carriers of the system remain un- In Fig. 3, we show the energy band structures and changed. Thus, it rules out the possibility of the effect theircorrespondingDOS.Bycomparingtheenergyband from the changes of DOS on superconductivity in our structures and DOS with different Pd dopants, the low system. energyfeaturesofelectronicstructuresremainalmostthe Another issue is the possible impurity scattering ef- same, except for lifting the degenerate bands and chang- fect [15], which comes from the Palladium substitution ing their dispersion at some high symmetric momentum for Platinum. This possibility can also be ruled out points around the Fermi level. This originates from the since SrPt P was reported to be an s-wave superconduc- nature of dopant Pd. When the atom Pt in SrPt P is 3 3 tor.[3,10]AccordingtotheAnderson’stheorem[16],the partially substituted by Pd, the high symmetry of point non-magneticimpuritydoesnotleadtoanapparentpair- groupofsystemwasloweredleadingtotheliftingofsome breaking effect in a conventional s-wave superconductor, degenerateenergybands. Inaddition,sincethePdatom andthusdoesnotsuppressthetransitiontemperatureT with 4d orbital has a stronger electron correlation than c apparently,whichisinsharpcontrasttothatinad-wave that of Pt with 5d orbital, the strong electron-electron superconductor, where the gap function has a nodal line interaction changes the low energy dispersion and makes and the zero energy excitation spectra can be modified it more flat with a large effective mass. significantly by non-magnetic impurities leading to the To quantitatively find the relation between electron 4 (a) x=0 (b) x=0.17 24 (d) x=0 (c) x=0.33 ) 20 x=0.17 V x=0.33 e s/ 16 e at St 12 ( S 8 O D 4 0 -6 -5 -4 -3 -2 -1 0 1 2 Energy (eV) FIG.3: (coloronline)TheelectronicbandstructurecalculationsforSr(Pt Pd ) Pwithdoping(a)x=0,(b)x=0.17,and 1−x x 3 (c) x = 0.33. (d) The corresponding density of states (DOS) to the three samples. The Fermi energy was set to zero (dashed lines). ity, χ ((cid:126)q), is given by: 0 1 (cid:88) |(cid:104)k+(cid:126)q,µ|k,ν(cid:105)|2 χ ((cid:126)q)= [f(E )−f(E )], 0 N E −E +i0+ ν,k µ,k+q(cid:126) k µ,k+q(cid:126) ν,k µνk where E represents the band energy measured at µ,k Fermi level E , and f(E ) is the Fermi-Dirac distri- F µ,k bution function for an eigenstate, |k,µ(cid:105). In addition, N k denotes the number of k points used for the irreducible Brillouin zone integration. The calculated real part of electron susceptibility is shown in Fig. 4, which demon- strates that the peak of electron susceptibility is sup- pressed as increase the Pd dopant concentrations. This resultalongwithourexperimentalfactssuggesttheelec- tron susceptibility of system originating from the itin- erant electrons with strong spin-orbit coupling, which is responsible for superconductivity, is weakened by intro- FIG.4: (coloronline)Therealpartofbareelectronsuscepti- ducing the electron correlation, and thus suggesting this bility χ ((cid:126)q) along the path between high symmetric momen- system displays a competition between the electron cor- 0 tum points. relation and spin-orbit coupling, which seems to be un- favorable for superconductivity. IV. CONCLUSION correlation and spin-orbit coupling in the present isova- lentPddopedsystem,wefurthercarriedouttheelectron In conclusion, we have successfully substituted Pd ele- susceptibilitycalculations. Thebareelectronsusceptibil- mentsintothepositionofPtinthestrongspin-orbitcou- 5 pling superconductor SrPt P, and found that the doping the “Strategic Priority Research Program (B)” of the 3 of Pd not only leads to the change of lattice parameters, Chinese Academy of Sciences (No. XDB04040300), and but also suppresses the superconducting transition tem- the Youth Innovation Promotion Association of the Chi- peratureT . Inaddition,thebandstructurecalculations nese Academy of Sciences (No. 2015187). This work c reveal that the calculated strength of electron suscepti- is partly sponsored by the Science and Technology Com- bility is suppressed as increase the Pd dopant concen- missionofShanghaiMunicipality(No. 14DZ2260700and trations. These results suggest that the competition be- 14521102800). tweenspin-orbitcouplingandelectroncorrelationplaysa vital role in superconductivity of the present 5d electron system. Acknowledgments ThisworkissupportedbytheNaturalScienceFounda- tion of China (No. 11204338, 11227902, and 11404359), [1] J. G. Bednorz and K. A. Mu¨ller, Possible high T super- nodelesssuperconductivityinthenoncentrosymmetricsu- c conductivity in the Ba-La-Cu-O system, Z. Phys. B 64, perconductor Mg Ir B , Phys. Rev. B 76, 064527 12−δ 19 16 189 (1986). (2007). [2] Y. Kamihara, T. Watanabe, M. Hirano and [12] C. Dong, PowderX: Windows-95-based program for pow- H. Hosono, Iron-based layered superconductor der X-ray diffraction data processing, J. Appl. Crystal- La[O F ]FeAs(x=0.05-0.12) with T = 26 K, J. logr. 32, 838 (1999). 1−x x c Am. Chem. Soc. 130, 3296 (2008). [13] F. Han, X. Zhu, P. Cheng, G. Mu, Y. Jia, L. Fang, Y. [3] T. Takayama, K. Kuwano, D. Hirai, Y. Katsura, A. Ya- Wang, H. Luo, B. Zeng, B. Shen, L. Shan, C. Ren and mamoto and H. Takagi, Strong coupling suerconductiv- H.H.Wen,SuperconductivityandPhasediagramsin4d- ity at 8.4K in an antiperovskite phosphide SrPt P,Phys. and 5d-metal-doped Iron Arsenides SrFe M As (M= 3 2−x x 2 Rev. Lett. 108, 237001 (2012). Rh, Ir, Pd), Phys. Rev. B 80, 024506 (2009). [4] C.-J. Kang, K.-H. Ahn, K.-W. Lee and B. I. Min, Elec- [14] Q.C.Ji,Y.H.Ma,K.K.Hu,B.Gao,G.Mu,W.Li,T. tron and Phonon Band-Structure Calculations for the Hu,G.H.Zhang,Q.B.Zhao,H.Zhang,F.Q.Huangand Antipolar SrPt P Antiperovskite Superconductor: Ev- X. M. Xie, Synthesis, Structural, and Transport Proper- 3 idence of Low-Energy Two-Dimensional Phonons, J. ties of Cr-Doped BaTi As O, Inorg. Chem. 53, 13089 2 2 Phys. Soc. Jpn. 82, 053703 (2013). (2014). [5] H. Chen, X. Xu, C. Cao and J. Dai, Charge Density [15] K. K. Hu, B. Gao, Q. C. Ji, Y. H. Ma, H. Zhang, G. Wave Instabiity and Soft Phonon in APt P(A=Ca, Sr, Mu, F. Q. Huang, C. B. Cai and X. X. Xie, Impurity 3 and La), Phys. Rev. B 86, 125116 (2012). ScatteringEffectinPd-dopedSuperconductorsofSrPt P, 3 [6] R. Szcz¸e´sniak, A. P. Durajski and L(cid:32). Herok, Theoretical submitted to Frontiers of Physics (2015). Description of the SrPt P Superconductor in the Strong- [16] P. W. Anderson, Theory of Dirty Superconductors, J. 3 Coupling Limit, Phys. Scr. 89, 125701 (2014). Phys. Chem. Solids 11, 26 (1959). [7] I. A. Nekrasov and M. V. Sadovskii, Electronic Struc- [17] G. Xiao, M. Z. Cieplak, J. Q. Xiao, and C. L. ture of New Multiple Band Pt-Pnictide Superconductors Chien, Magnetic pair-breaking effects: Moment for- APt P, JETP Lett. 96, 227 (2012). mation and critical doping level in superconducting 3 [8] B. I. Jawdat, B. Lv, X. Zhu, Y. Xue and C.-W. Chu, La Sr Cu A 0 systems (A = Fe, Co, Ni, Zn, 1.85 0.15 1−x x 4 High-pressureandDopingStudiesoftheSuperconducting Ga, Al) , Phys. Rev. B 42, 8752 (1990). Antiperovskite SrPt P, Phys. Rev. B 91, 094514 (2015). [18] W. Li, X. Y. Wei, J. X. Zhu, C. S. Ting and Y. Chen, 3 [9] D. A. Zocco, S. Krannich, R. Heid, K.-P. Bohnen, T. Pressure Induced Topological Quantum Phase Transition Wolf, T. Forrest, A. Bossak and F. Weber, Lattice Dy- in Sb Se , Phys. Rev. B 89, 035101 (2014). 2 3 namicalPropertiesofSuerconductingSrPt PStudiedvia [19] G.KresseandJ.Furthmuller,EfficientIterativeSchemes 3 Inelastic X-Ray Scattering and Density Functional Per- for ab Initio Total-Energy Calculations Using a Plane- turbation Theory, arXiv:1510.02012 (2015). Wave Basis Set, Phys. Rev. B 54, 11169 (1996). [10] T. Shiroka, M. Pikulski, N. D. Zhigadlo, B. Batlogg, J. [20] J. P. Perdew, K. Burke and M. Ernzerhof, Generalized MesotandH.-R.Ott,Paring of Weakly Correlated Elec- Gradient Approximation Made Simple, Phys. Rev. Lett. trons in the Platinum-Based Centrosymmetric Supercon- 77, 3865 (1996). ductor SrPt P, Phys. Rev. B 91, 245143 (2015). 3 [11] G. Mu, Y. Wang, L. Shan, and H. H. Wen, Possible