Scanning Tunneling Microscopy in the superconductor LaSb2
J.A. Galvis, H. Suderow, S. Vieira, S.L. Bud'ko, P.C. Canfield
SScanning Tunneling Microscopy in the superconductor LaSb J. A. Galvis,
1, 2, 3
H. Suderow ∗ ,
1, 2, 3
S. Vieira,
1, 2, 3
S. L. Bud’ko, and P. C. Canfield Laboratorio de Bajas Temperaturas, Departamento de F´ısica de la Materia Condensada,Instituto de Ciencia de Materiales Nicol´as Cabrera, Facultad de Ciencias,Universidad Aut´onoma de Madrid, E-28049 Madrid, Spain Condensed Matter Physics Center (IFIMAC), Universidad Aut´onoma de Madrid, E-28049 Madrid, Spain Unidad Asociada de Bajas Temperaturas y Altos Campos Magn´eticos,UAM/CSIC, Cantoblanco, E-28049 Madrid, Spain Ames Laboratory and Department of Physics and Astronomy, Iowa State University, Ames, IA50011, USA (Dated: October 29, 2018)We present very low temperature (0.15 K) scanning tunneling microscopy and spectroscopy ex-periments in the layered superconductor LaSb . We obtain topographic microscopy images withsurfaces showing hexagonal and square atomic size patterns, and observe in the tunneling conduc-tance a superconducting gap. We find well defined quasiparticle peaks located at a bias voltagecomparable to the weak coupling s-wave BCS expected gap value (0.17 meV). The amount of statesat the Fermi level is however large and the curves are significantly broadened. We find T c of 1.2 Kby following the tunneling conductance with temperature. PACS numbers: 74.50.+r, 74.62.-c, 74.25Jb
INTRODUCTION
LaSb is a layered rare-earth diantimonide which crys-tallizes in the layered orthorhombic SmSb structure[1–3]. La/Sb bi-layers of triangular prims alternate with twodimensional (2D) sheets of Sb as shown in Fig.1. Theisostructural rare-earth diantimonides RSb (R=La-Nd,Sm) show highly anisotropic transport properties[2]. Re-sistivity measurements in LaSb show that the tempera-ture dependence is stronger for a current applied parallelto the c-axis than for a current applied in-plane[2, 3].On the other hand, susceptibility measurements showthat LaSb is a weakly temperature-dependent isotropicdiamagnet[2, 3]. In-plane magnetoresistance measure-ments at low temperatures (2 K) give an anisotropy forin-plane and c-axis applied magnetic fields that is verypronounced. Moreover, the in-plane magnetoresistanceis linear and does not saturate up to 45 T. LaSb hasbeen proposed as a good candidate to make high mag-netic field sensors[4]. The observed linear magnetoresis-tance is at odds with usual magnetic field dependenceof the magnetoresistance[7]. In metals, the magnetore-sistance can saturate when Fermi surface is closed; re-main quadratic up to high fields and then saturate whenFermi surface is open; or it can continously evolve eventu-ally showing quantum oscillations in compensated metalswith equal number of electrons and holes[5, 6].On the other hand, at higher temperatures, the mag-netoresistance exhibits a pronounced maximum above 20K, whose origin is yet unknown[2]. As there are no ev-idences for magnetic correlations in this material, thepossible relationship to some sort of magnetic field in-duced charge density wave (CDW) has been highlightedin Ref.[2]. This possibility has been, in turn, used totry to explain the low temperature linear dependence C LaSb
La/Sb bilayerSb sheet
FIG. 1: Lattice structure of LaSb . A top view of one La/Sbbi-layers and one sheet of Sb is shown. Lattice parametersare a = 6 . A , b = 6 . A and c = 18 . A . of the magnetoresistance[4, 8]. Yet another possibleexplanation is a linear energy spectrum[9]. The com-pound LaAgSb , which also shows a large positive mag-netoresistance up to 18 T[10, 11], exhibits an anomalyin the resistance around 210 K which was associated toa CDW transition[10, 12, 13]. It has been noted thatother layered materials such as 2H-NbSe and 2H-TaSe ,where CDWs open at, respectively, 30 K and 120 K,also show roughly linear magnetoresistance[14, 15]. Inspite of all these hints, until now there is no evidence forCDW in LaSb , including most recent optical conductiv-ity measurements[3]. a r X i v : . [ c ond - m a t . s up r- c on ] J un The Fermi surface has been measured using de Haas-van Alphen oscillations in Ref.[8]. There are threeFermi surface sheets, two of them nearly two-dimensional(cylindrical along c), and one nearly spherical. All ofthem are very small, pointing out that LaSb has a smallnumber of carriers. Band structure calculations give fur-ther additional two bands, yet unobserved in de Haasvan Alphen. Photoemission shows that Sb 5p states, hy-bridized with La 5d states, dominate the Fermi surfacefeatures, and that electronic properties are rather two-dimensional[18].Among the rare earth diantimonides, LaSb is uniquein that it shows a superconducting state. Superconduct-ing features are not clearly established in this material.Early reports give a very low T c of 0.4 K[19]. However,more recent electrical transport and magnetic suscepti-bility measurements show a very broad transition, witha fragile onset as high as 2.5 K[3, 20]. The resistanceevolves then to roughly zero around 0.4 K-0.5 K. The su-perconducting properties exhibit strong dependence withthe pressure, which reduces the anisotropy and sharpensthe transition, with a T c = 2.1 K at 4 kbar[20]. Here weprovide scanning tunneling microscopy and spectroscopy(STM/S) measurements in the superconducting phaseof LaSb . We observe the atomic lattice at tempera-tures down to 150 mK. We find a critical temperature ofT c =1.2 K, which is within the range of the broad transi-tion found in macroscopic measurements[3, 20]. EXPERIMENT
We use a homebuilt STM/S system installed in a di-lution refrigerator with an energy resolution in the tun-neling spectroscopy of 0.15 K. This system is inserted ina vector magnetic field solenoid and has a similar con-struction to the one described in Ref.[21]. We use a tipprepared from a gold wire cut with a clean blade, which iscleaned in-situ[22]. Single crystals of LaSb were grownfrom high-purity La and Sb in excess of Sb flux[23–25].Large residual resistance ratios are found, of about 60, asin Ref.[2]. The crystals grow as plates with the c-axis per-pendicular to the plates. They consist of soft Aluminium-foil like sheets that can be peeled off by glueing a stickand pushing it. We did not find good tunneling condi-tions in surfaces obtained immediately after an in-situlow-temperature cleave. After heating to room tempera-ture, we realized that this was due to loose sheets remain-ing on the surface giving bad tunneling conditions. Bycleaving the sample at ambient and removing those loosesheets using tweezers, we were able to obtain opticallyflat and shiny surfaces with no loose sheets. We measuredtwo different samples, studying about four scanning win-dows in different regions of sample. Of course, surfacecontamination cannot be totally avoided when preparingthe sample at ambient conditions. However we were able to find good scanning conditions at low temperatures,with reproducible imaging, being independent of the tun-neling conductance. In the topographic images, the tip-sample bias voltage is fixed at 2 mV and the tunnelingcurrent remains constant while the tip is scanned overthe sample surface at a tunneling conductance of orderof or below a µ S. The conductance vs bias voltage curvesare obtained, as usual, by cutting the feedback loop andnumerically differentiating the obtained current vs volt-age curves[21, 26]. Tunneling conductance is normalizedto one at bias voltages above one mV.
RESULTS
In Fig.2 we show atomic resolution features in the to-pography images taken on different surfaces of the sam-ple at 0.15 K. We can find hexagonal and also squareatomic lattice arrangements (Fig.1). The obtained ge-ometry is in good agreement with the cuts on the struc-ture shown in the right panels of Fig.1. The lattice pa-rameters also coincide with those measured with X-rayscattering[4]. Accordingly, the surfaces showing squareshapes are probably made out of Sb atoms and of La inthose showing hexagonal shapes. -1 -1 ab FIG. 2: In the left panels we show atomic resolution topog-raphy images at 0.15 K on two different LaSb surfaces, with(a) hexagonal and (b) square atomic sized shapes. These im-ages are unfiltered. The corresponding Fourier transforms isgiven in the right panels. Middle panels are Fourier filteredimages, made by filtering out everything except green circles,which show the position of the Bragg peaks due to the atomicmodulation.[27] Fig.3a shows the tunneling conductance vs bias volt-age and its temperature dependence. We find supercon-ducting features with rounded quasi-particle peaks anda large conductance at zero and low bias. We couldnot observe clear atomic size variations of the supercon-ducting features, as those previously measured in other -0.4 -0.2 0.0 0.2 0.40.40.81.21.62.02.42.83.2 N o r m a li z e d t unn e li ng c ondu c t a n ce Bias voltage (mV)
T/T C ( ) / ( ) ba -0.4 -0.2 0.0 0.2 0.4 Bias voltage (mV) N o r m a li z e d c ondu c t a n ce FIG. 3: a) Temperature dependence of the experimental tun-neling conductance (black curves) and calculated conductance(red line). The data are, from bottom to top, taken at 0.15,0.27, 0.36, 0.48, 0.55, 0.65, 0.77, 0.82, 0.92, 1.1 and 1.2 K. b)Temperature dependence of the position of the quasiparticlepeak in the density of states used to calculate the red linesin a. The black solid line is the temperature dependence ofthe superconducting gap in BCS theory. Inset shows againthe lowest temperature curve in a. Blue dashed lines givequasiparticle peak positions for a weak coupling BCS super-conductor with T c of 1.2 K. compounds[16, 17, 29]. The superconducting featuresdisappear fully into a flat tunneling conductance at about1.2 K. At the lowest temperatures, the tunneling conduc-tance curves show rounded quasi-particle peaks whosemaximum is located at 0.17 mV. This can be fitted (redline in inset of Fig.3b) to a density of states showing alarge amount of states at the Fermi level (25% of thehigh bias voltage conductance) and a distribution of val-ues of the superconducting gap, which ranges from 0.02meV to 0.2 meV. The density of states and tunnelingconductances are calculated by simply introducing a gapdistribution in the BCS density of states formula, asmade earlier in several superconductors showing multi-band properties [16, 30, 31]). The value for the super-conducting gap ∆ expected from the weak coupling equa-tion ∆ = 1 . k B T c for T c = 1 . K , gives ∆ =0.182 meV(highlighted with the dotted blue line in Fig.3b). When -0.4 -0.2 0.0 0.2 0.40.00.20.40.60.81.01.21.41.61.82.0 Bias voltage (mV) N o r m a li z e d t unn e li ng c ondu c t a n ce B ab T B ll ab V = Magnetic Field (mT) V = Magnetic Field (mT)
FIG. 4: Magnetic field dependence of tunneling conductancecurves at 0.15 K for magnetic field applied perpendicular (a)and parallel (b) to the ab-plane. Insets show the magneticfield dependence of the zero bias conductance σ ( V = 0 mV ). we increase the temperature, the tunneling conductancecurves can be fitted by a density of states similar to theone at lowest temperatures, but with the gap distributiongradually shifted to lower energies when increasing tem-perature. The temperature dependence of the quasipar-ticle peak position in the corresponding density of statesis shown in Fig.3b, and is close to but somewhat belowthe BCS temperature dependence for the gap (black linein Fig.3b).The magnetic field dependence of the tunneling con-ductance is shown in top and bottom panels of Fig. 4with the field applied perpendicular and parallel to theplane. The curves are made following the same methodas in Ref.[32]. We modify the scanning window where thecurves are taken on the surface to find the largest value ofthe gap over the sample. The magnetic field leads to anadditional broadening and flat curves above the criticalfield. We find an anisotropic critical field with a factor ofapproximately 4 between the parallel and perpendicularfield. Using H c ⊥ = φ / πξ ⊥ and H c (cid:107) = φ / πξ ⊥ ξ (cid:107) wefind in-plane and c-axis coherence lengths, ξ ⊥ ∼
46 nmand ξ (cid:107) ∼
10 nm. Such a high value of in-plane coherencelength ξ ⊥ requires rather large flat areas comprising sev-eral 100 nm to be able to see vortices. We did not obtainsuch large flat areas in our experiment. DISCUSSION AND CONCLUSIONS
First, let us remark that the atomic resolution topogra-phy images and its associated Fourier transform (see Fig.2) do not show any trace of modulations different to theatomic periodicity. Therefore, we find no evidence forCDW. This does not totally rule out its existence, butlikely the CDW corresponds either to large wavelengthmodulations of the atomic lattice, above the size of ouratomic images (roughly 10 lattice constants), or it occursalong the c-axis. The tunneling conductance up to 5 mV,which is the range we have explored, is also flat, showingthat there is no pseudogap like behavior in this energyrange. This is relevant, because the wide resistive tran-sition could be related pre-formed pairs or some sort ofpseudogap behavior as in the cuprates[29]. The featuresobserved here at the surface of LaSb are instead thoseof a good metal.Regarding superconductivity, the shape of the tun-neling conductance curves measured here present strongbroadening. There is a distribution of values of the su-perconducting gap and a large amount of states close orat the Fermi level. The critical temperature is well de-fined and relatively high as compared to other measure-ments, T c = 1 . , which shows an extremely small bulk critical tem-perature of 0.15 K[17]. On compressing layers by apply-ing pressure, T c increases rapidly to above 1 K with somekbar [37]. In 2H-TaSe , STM measurements show signif-icantly increased critical temperature (up to 1 K) withrespect to the bulk[17]. At the same time, broadened gapfeatures appear instead of the fully gapped superconduc-tivity observed in similar materials such as 2H-NbSe [28]. In LaSb the surface superconducting properties havehigher T c than the resistive transition and we also ob-serve broadened superconducting tunneling conductancefeatures. The comparison with the dichalchogenides sug-gests that surface superconducting properties can bemore robust and well defined than the bulk properties inlayered materials. The last layers have possibly relaxedinternal strains present in the bulk, eliminating inhomo-geneous behavior observed in macroscopic properties.We should note that, although the last layers show asingle clean transition, the high amount of states closeto the Fermi level points to anomalous superconductingproperties. Its origin remains unclear, and they could berelated to pair breaking effects due to disorder or couplingto nearby layers or regions with different superconduct-ing properties. The broadened quasiparticle peaks showa wide distribution of values of the superconducting gap,mostly below the expected BCS value. This is compatiblewith multigap scenarios observed in two-band supercon-ductors in layered materials[16, 31, 33–36]. Multiband ormultigap superconductivity can be expected within theFermi surface features of LaSb [8].The appearance of stronger superconductivity at thesurface suggests on the other hand that few or single layersystems could show superconducting properties. The fab-rication of few layer sheets of LaSb and similar com-pounds should be feasible, as they are easily peeled offand have a shape which is very similar to transition metaldichalchogenides. In the latter case, single layers havebeen synthesized already, showing in some cases super-conducting correlations[17, 38–40]. LaSb belongs to thefamily of RSb , where R is a rare earth[2]. All compoundsof the series appear in form of similarly shaped layeredcrystals. Having a rare earth in the crystal structure,they can lead to further interesting superconducting andmagnetic properties not found in transition metal dichal-chogenide layers.In summary, we have observed the crystalline structureof LaSb by means of atomic resolution STM measure-ments. The atomic lattice images at 0.15 K do not showsignatures of charge density wave. In the tunnelling con-ductance curves, we observe broadened superconductingfeatures disappearing at 1.2 K. The observed behaviorcan be related to the layered structure. We would liketo highlight the remarkable combination of high puritysamples, as seen in the residual resistance ratio, and su-perconducting features that are more robust and betterdefined at the surface.This work was supported by the Spanish MINECO(Consolider Ingenio Molecular Nanoscience CSD2007-00010 program, FIS2011-23488, ACI-2009-0905), by theComunidad de Madrid through program Nanobiomagnetand by COST MP1201. Work at the Ames Laboratorywas supported by the US Department of Energy, BasicEnergy Sciences, Division of Materials Sciences and En-gineering under Contract No. DE-AC02-07CH11358. ∗ Corresponding author, [email protected]. [1] R. Wang and H. Steinfink, Inorg. Chem. , 1685 (1967).[2] S. L. Bud’ko, P. C. Canfield, C. Milelke, and A. H. Lac-erda, Phys. Rev. B. , 13624 (1998).[3] J. F. DiTusa and et al, J. Phys.: Conf. Ser. , 012151(2011).[4] R. Young, R. G. Goodrich, J. F. DiTusa, S. Guo, P. W.Adams, and et al., Appl. Phys. Lett. , 21 (2003).[5] P. B. Alers and R. T. Webber, Physical Review B ,1060 (1953).[6] H. Yang, B. Fauque, L. Malone, A. Antunes, Z. Zhu,C. Uher, and K. Behnia, Nat. Comm. , 47 (2010).[7] A. A. Abrikosov, J. Phys. A: Mathematical and General , 9119 (2003).[8] R. G. Goodrich, D. Browne, R. Kurtz, D. P. Young, J. F.DiTusa, P. W. Adams, and D. Hall, Phys. Rev. B ,125114 (2004).[9] A. A. Abrikosov, Phys. Rev. B , 4231 (1999).[10] K. Myers, S. Bud’ko, I. Fisher, Z. Islam, H. Kleinke,A. Lacerda, and P. C. Canfield, J. Magn. Magn. Mater. , 27 (1999).[11] K. Wang and C. Petrovic, Phys. Rev. B , 155213(2012).[12] C. Song, J. Park, J. Koo, K. B. Lee, J. Y. Rhee, S. L.Bud’ko, P. C. Canfield, B. N. Harmon, and A. I. Gold-man, Phys. Rev. B , 035113 (2003).[13] C. Lue, Y. Tao, K. Sivakumar, and Y. Kuo, J. Phys.Cond. Matt. , 406230 (2007).[14] J. A. Wilson, F. J. DiSalvo, and S. Mahajan, Adv. Phys. , 117 (1975).[15] P. Monceau, Adv. Phys. , 325 (2012).[16] I. Guillam´on, H. Suderow, S. Vieira, L. Cario, P. Diener,and P. Rodi`ere, Phys. Rev. Lett. , 166407 (2008).[17] J. A. Galvis, P. Rodi´ere, I. Guillam´on, M. R. Osorio,J. G. Rodrigo, L. Cario, E. Navarro-Moratalla, E. Coron-ado, S. Vieira, and H. Suderow, Phys. Rev. B , 094502(2013).[18] A. I. Acatrinei, D. Browne, Y. B. Losovyi, D. P. Young,M. Moldovan, J. Y. Chan, and R. L. Kurtz, Phys. Con-dens. Matter , 33 (2003).[19] F. Hullinger and H. R. Ott, Less-Common Met , 103 (1977).[20] S. Guo, D. P. Young, P. W. Adams, X. S. Wu, J. Y.Chan, G. T. McCandless, and J. F. DiTusa, Phys. Rev.B , 174520 (2011).[21] H. Suderow, I. Guillamon, and S. Vieira, Rev. Sci. Inst. , 033711 (2011).[22] J. G. Rodrigo, H. Suderow, S. Vieira, E. Bascones, andF. Guinea., J. Phys.: Condens. Matter (2004).[23] P. C. Canfield and Z. Fisk, Phil. Mag. B , 1117 (1992).[24] P. C. Canfield, 2nd Euroschool on Complex Metals , 1(2005).[25] P. C. Canfield, J. D. Thompson, and Z. Fisk, J. Appl.Phys. , 5992 (1991).[26] H. Suderow, P. Martinez-Samper, S. Vieira, N. Luchier,J. P. Brison, and P. C. Canfield, Phys. Rev. B , 020503(2001).[27] I. Horcas, R. Fernandez, J. M. Gomez-Rodriguez,J. Colchero, J. Gomez-Herrero, and A. M. Baro, Rev.Sci. Instrum. , 013705 (2007).[28] I. Guillamon, H. Suderow, F. Guinea, and S. Vieira,Phys. Rev. B , 134505 (2008).[29] O. Fischer, M. Kugler, I. Maggio-Aprile, C. Berthod, andC. Renner, Rev. Mod. Phys. , 353 (2007).[30] M. Crespo, H. Suderow, S. Vieira, S. Bud’ko, andP.C. Canfield, Phys. Rev. Lett. , 027003 (2006).[31] J. G. Rodrigo, and S. Vieira, Physica C , 306 (2004).[32] J. A. Galvis, M. Crespo, I. Guillamon, H. Suderow,S. Vieira, M. G. Hernandez, S. Bud’ko, and P. C. Can-field, Sol. State Comm. , 1076 (2012).[33] E. Bascones, and F. Guinea, Phys. Rev. B , 214508(2001).[34] M. Zehetmayer, Superc. Sci. Tech. , 043001 (2013).[35] A. Y. Liu, I. I. Mazin, and J. Kortus, Phys. Rev. Lett. , 087005 (2001).[36] G. Rubio, H. Suderow, and S. Vieira, Phys. Rev. Lett. , 5582 (2001).[37] T. Smith, R. Shelton, and R. Schwall, J. Phys. F: MetalPhys. , 1713 (1975).[38] A. Castellanos-Gomez, N. Agrait, and G. Rubio-Bollinger, Applied Physics Letters , 213116 (2010).[39] N. E. Staley, J. Wu, P. Eklund, Y. Liu, L. Li, and Z. Xu,Phys. Rev. B , 184505 (2009).[40] J. Ye, Y. Zhang, M. Bahramy, R. Arita, and Y. Iwasa,Science338