Generic Fe buffer layers for Fe-based superconductors: Epitaxial FeSe1-xTex thin films
Kazumasa Iida, Jens Haenisch, Michael Schulze, Saicharan Aswartham, Sabine Wurmehl, Bernd Buechner, Ludwig Schultz, Bernhard Holzapfel
GGeneric Fe buffer layers for Fe-based superconductors: Epitaxial FeSe − x Te x thin films Kazumasa Iida, a) Jens H¨anisch, Michael Schulze, Saicharan Aswartham, Sabine Wurmehl, Bernd B¨uchner, Ludwig Schultz, and Bernhard Holzapfel Institute for Metallic Materials, IFW Dresden, D-01171 Dresden, Germany Institute for Solid State Research, IFW Dresden, D-01171 Dresden, Germany (Dated: 8 November 2018)
Biaxially textured FeSe − x Te x films have been realized on Fe-buffered MgO substrates by pulsed laser de-position. Similar to the Fe/BaFe As bilayers, the crystalline quality of FeSe − x Te x films exhibit a sharpout-of-plane and in-plane texture less than 0.9 ◦ . The Fe/FeSe − x Te x bilayers showed high superconductingtransition temperatures of over 17 K. The angular-dependent critical current densities exhibit peaks posi-tioned at H ⊥ c similar to other pnictide thin films. The volume pinning force of FeSe − x Te x in this directionis very strong compared with that of Co-doped BaFe As , due to a good matching between the interlayerdistance in the c direction and the out-of-plane coherence length.PACS numbers: 74.70.Xa, 81.15.Fg, 74.78.-w, 74.25.Sv, 74.25.F-Epitaxial iron chalcogenide superconducting thin filmshave been prepared on single crystalline oxides substratesby pulsed laser deposition (PLD) or molecular beamepitaxy (MBE) to investigate their intrinsic propertiesand to explore superconducting device applications. These superconductors are very sensitive to strain. In-deed, the superconducting transition temperature ( T c )of FeSe . Te . films can be tuned by strain and evenhigher T c values in films than in bulk materials can berealized. In addition, the superconductivity is induced inFeTe films by the tensile stress albeit their bulk crystalsare not superconducting at ambient pressure. The biaxial strain is usually induced by a lattice mis-match between films and substrates. However, the corre-lation between T c and lattice mismatch for FeSe . Te . is controversial. A fundamental problem is the forma-tion of an interfacial layer between FeSe . Te . and ox-ide substrates, which compromises the epitaxial growth.Later the interfacial layer was identified as amorphouscontaining oxygen, which comes mainly from oxidesubstrates. Furthermore, oxygen diffusion into the filmswas detected by transmission electron microscope (TEM)analyses, which deteriorates the superconducting proper-ties. In order to minimize the oxygen diffusion, Tsukada et al . have proposed the implementation of CaF sub-strates, resulting in good superconducting properties. Hence, ideal growth conditions of FeSe − x Te x thin filmsrequire non-oxide substrates under ultra-high vacuum(UHV) atmosphere.Recently we have reported that epitaxial Co-dopedBaFe As (Ba-122) films can be realized on Fe-bufferedsingle crystalline MgO substrates in UHV condition. The detailed TEM analyses revealed that the FeAs tetra-hedron in the Ba-122 bonds coherently to body cen-tered cubic Fe. This Fe/Ba-122 bilayer has a clean mi- a) Electronical address: [email protected] (b)(a)
10 nm0 nm
FIG. 1. (Color online) (a) The RHEED image of FeSe − x Te x at room temperature shows only streaks, indicative of a flatsurface. The incident electron beam is the MgO [110] az-imuth. (b) The corresponding AFM image (1 µ m × µ m) showsan R rms value of 0.73 nm. crostructure and is of excellent crystalline quality with-out grain boundaries (GBs). Since the FeAs or FeSe(Te)tetrahedron is a common structure for both Ba-122 andFeSe − x Te x , the implementation of an Fe buffer shouldbe also applicable for the epitaxial growth of FeSe − x Te x .In this letter, we demonstrate the implementation of anFe buffer layer to grow epitaxial FeSe − x Te x films andpresent their transport properties.The Fe/FeSe − x Te x bilayers were prepared by PLD(KrF excimer laser, λ = 248 nm), which is almost iden-tical deposition methods to Fe/Ba-122 except for thedeposition temperature and the laser repetition rate ofFeSe − x Te x . The energy density of the laser on thetarget was 3-5 J/cm and the distance between targetand substrate was approximately 6 cm. Single crystallineMgO (100) substrates were heated to 1000 ◦ C, held atthis temperature for 30 min, and subsequently cooledto room temperature for cleaning. After the Fe depo-sition at room temperature, the Fe-covered MgO washeated to 750 ◦ C, held at this temperature for 20 min andsubsequently cooled to 450 ◦ C, the optimum depositiontemperature of FeSe − x Te x films with highest T c value,similarly to the results in Ref. . Once the temperature a r X i v : . [ c ond - m a t . s up r- c on ] N ov was stabilized, the FeSe − x Te x layer was deposited at alaser frequency of 3 Hz. The whole deposition processwas conducted under UHV condition (base pressure of10 − mbar). The respective layer thickness of Fe andFeSe − x Te x were 20 nm and 95 nm confirmed by cross-sectional focused ion beam cuts at different sample areasas well as TEM.The PLD target was prepared by a modified Bridgmantechnique yielding an Fe-Te-Se crystal with the nominalcomposition of Fe:Se:Te=1:0.5:0.5. For the growth, sto-ichiometric amounts of pre-purified metals were sealedin an evacuated quartz tube. The tube was placed ina horizontal tube furnace and heated up to 650 ◦ C andkept at that temperature for 24 h. The furnace was thenheated to 950 ◦ C and the temperature was kept constantfor 48 h. Finally, the furnace was cooled down with arate of 5 ◦ C/h to 770 ◦ C, followed by furnace cooling.We yield crystals with dimensions up to cm-size. A bulk T c of 13.6 K was recorded by a superconducting quan-tum interface device (SQUID) magnetometer. Details ofthe single crystal preparation and their properties can befound in Ref. .Each deposition step was monitored by reflection high-energy electron diffraction (RHEED). The RHEED im-age of the FeSe − x Te x acquired at room temperature,fig. 1 (a), shows only streaks similarly to the Fe bufferlayer. This is indicative of a smooth surface of theFeSe − x Te x layer, which is consistent with the obser-vation by atomic force microscope (AFM) presentedin fig. 1 (b). A root mean square roughness ( R rms ) of0.73 nm was recorded.In order to check phase purity and texture qualityof the films, detailed structural characterizations by x-ray diffraction were conducted as summarized in fig. 2.The θ/ θ - scans, fig. 2 (a), show only 00 l reflections ofFeSe − x Te x together with the 002 reflection of Fe andMgO, indicating c -axis texture and phase purity. The ω - scan for the 001 reflection of FeSe − x Te x in fig. 2 (b)shows a sharp full width at half maximum (FWHM, ∆ ω )of 0.72 ◦ . The 101 pole figure measurements and the cor-responding φ -scans of FeSe − x Te x reveal no satellite andadditional reflections other than sharp and strong reflec-tions at every 90 ◦ , indicative of biaxial texture. Here, theepitaxial relation is identified as (001)[100]FeSe − x Te x (cid:107) (001)[110]Fe (cid:107) (001)[100]MgO. The average ∆ φ of Fe andFeSe − x Te x are 1.05 ◦ and 0.83 ◦ , respectively. From theseresults, the FeSe − x Te x layer is of good crystalline qual-ity without GBs.In order to avoid any damage during ion beam etchingand to ensure a low contact resistance, an Au layer wasdeposited on the film at room temperature by PLD. TheAu-covered FeSe − x Te x film was then ion beam etchedto form bridges with 0.5 mm width and 1 mm lengthfor transport measurements, which were conducted in aPhysical Property Measurement System (PPMS, Quan-tum Design) by a four-probe method. Here, a criterionof 1 µ Vcm − ( E c ) was used for evaluating J c .The FeSe − x Te x film exhibits no sign of a resistance I n t e n s it y ( x10 c p s ) -2 0 2w (¡)(b)Dw=0.72 ¡ I n t e n s it y ( a r b . un it ) Te x FeMgO Df=0.83¡10 I n t e n s it y ( c p s )
001 002 003 004 F e , M g O , Kb(a) M g O , ( l / ) (c) FIG. 2. (Color online) (a) The θ/ θ - scan of an FeSe − x Te x bilayer in Bragg-Brentano geometry using Co-K α radiation.Both layers of Fe and FeSe − x Te x were grown with c -axistexture. (b) Rocking curve of the 001 reflection. ∆ ω = 0 . ◦ indicates good out-of-plane texture. (c) The 101 pole figuremeasurement and the corresponding φ -scans of FeSe − x Te x .The respective φ -scans of the 110 Fe and the 220 MgO arealso presented. The average ∆ φ of Fe and FeSe − x Te x are1.05 ◦ and 0.83 ◦ , respectively. m H i rr ( T ) c ) n , n = 1.33 (1-T/T c ) n , n = 0.77 (c)181716 T (K) r c H ^ c 0 T9 T(b)2.52.01.51.00.50.0 r ( x10 - m W c m ) r c H || c 0 T9 T(a)
FIG. 3. (Color online) (a) Resistivity traces for FeSe − x Te x film measured in several magnetic fields for H (cid:107) c and (b) H ⊥ c . Field increment is 1 T. Broken lines indicate the ρ c for µ H irr .(c) µ H irr for H (cid:107) c are always lower than that for H ⊥ c . Open symbols are evaluated by the Kramer extrapo-lation. anomaly at around 200 K, which has been frequently ob-served in FeSe − x Te x thin films. In addition, relativelysmall resistivity and a high residual resistivity ratio of 3.5is recorded. This is a consequence of the current shuntingeffect since the resultant film was fully covered with Au.Therefore, the normal state behavior is masked by Au.Nevertheless, the onset T c is recorded at 17.7 K, which ishigher than the bulk value. The higher T c might be dueto the strain effect as reported in Ref. . When magneticfields are applied to the film, a clear shift of T c to lowertemperatures is observed for both directions, as exhibitedin figs. 3 (a) and (b). In particular, this shift togetherwith a broadening of the transition is more significantfor H (cid:107) c than H ⊥ c . Figure 3 (c) shows the temperaturedependence of the irreversibility field ( H irr ), which is de-fined by the intersection between the resistivity tracesand the resistivity criterion ( ρ c ) as shown in figs. 3 (a)and (b). In addition, µ H irr evaluated by the Kramerextrapolation from J c − H characteristics presented laterare also plotted. Here the ρ c is defined as E c / J c , ,where the J c , = 15 Acm − is a criterion for H irr in J c − H characteristics. In this definition, the N -value ofthe voltage-current characteristics ( V − I , V ∝ I N ) is 1in the vicinity of E c . It is clear from fig. 3 (c) that µ H irr for H (cid:107) c is always lower than that for H ⊥ c , indicat-ing that the flux pinning is anisotropic. The µ H irr for H (cid:107) c shows a power law relation, µ H irr ∼ (1 − T /T c ) n ,with exponent n = 1 . ± .
02 in the temperature rangefrom 8 K to 17 K. At 4.2 K, an increased value for µ H irr is observed presumably due to multi-band effects. Onthe other hand, an exponent of n = (0 . ± . < H ⊥ c , which is very similar to the uppercritical field ( µ H c2 ) for H ⊥ c in layered compounds (i.e. µ H c2 ∝ (1 − T /T c ) . ). Shown in fig. 4 (a) are J c − H characteristics measuredfor H (cid:107) c at several temperatures. J c values are quicklyreducing with increasing H , indicative of less flux pinningin this direction. It should be noted that V − I curvesfor evaluating J c always show a power-law relation, sug-gesting that current is limited by depinning of flux ratherthan GBs. On the other hand, the reduction of J c with H is very small for H ⊥ c below 16 K (fig. 4 (b)). TheSe(Te)-Se(Te) interlayer distance in the c direction is al-most identical to the out-of-plane coherence length at lowtemperatures ( ξ c =0.35 nm), resulting in strong pinning inthis direction. Here the Se(Te)-Se(Te) interlayer distanceis calculated as around 0.25 nm using c (1 − z ) with thelattice parameter c and the Se(Te) coordination z . ForCo-doped Ba-122, the As-As interlayer distance in the c direction is 0.3 nm, which is far below ξ c (1.2 nm). As aresult, J c at H ⊥ c is decreased significantly with increas-ing H compared with FeSe − x Te x . The volume pinningforce ( F p ) of FeSe − x Te x is larger than that of Co-dopedBa-122 for both major directions at reduced tempera-tures of t = 0 .
68 and 0.78 (fig. 4 (c) and (d)). However, F p of Co-doped Ba-122 for H (cid:107) c is larger than that ofFeSe − x Te x at t = 0 .
48 (not shown in this letter). Theorigin of this observation has to be investigated in detail J c ( A / c m ) H (T) 4.2K8K10K12K14K16K (a)H || c J c,15 H (T)
H ^ c (b)J c,15 F p ( GN / m ) H (T) H || cFeSe Te x t=0.68t=0.78Ba-122t=0.68t=0.78 (c) 1086420 F p ( GN / m ) H (T) Ba-122FeSe Te x H ^ c (d)
FIG. 4. (Color online) (a) J c − H characteristics measuredin H (cid:107) c and (b) H ⊥ c at several temperatures. Chain linesindicate a J c criteria of 15 Acm − for µ H irr . (c) A com-parison of the F p between FeSe − x Te x and Co-doped Ba-122( T c = 23 . t = 0 .
68 and 0.78for H (cid:107) c and (d) H ⊥ c . The data of Co-doped Ba-122 plottedas lines are taken from the Ref. . J c ( A / c m ) FIG. 5. (Color online) J c (Θ) of the FeSe − x Te x film measuredin several magnetic fields at 8 K. No additional peaks exceptat H ⊥ c are observed. by TEM in future.The angular-dependent critical current densities( J c (Θ)) at 8 K in several magnetic fields, fig. 5, alwaysshow a peak positioned at Θ = 90 ◦ and 270 ◦ owing to theintrinsic pinning. Here the magnetic field was applied inthe maximum Lorentz force configuration ( H ⊥ J) at anangle Θ measured from the c -axis. A small J c anisotropy( γ J = J c (90 ◦ ) /J c (180 ◦ )) of 2.6 is observed at 1 T. This γ J is significantly increasing with H (e.g. γ J = 46 . µ H = 9 T) due to the relatively close to H irr in the c direction.In summary, epitaxial FeSe − x Te x films with sharpout-of-plane and in-plane texture have been realized onFe-buffered single crystalline MgO substrates similar tothe Fe/Ba-122 bilayers. These results indicate that Fecan work as generic buffer layer for epitaxial growth ofFe-based superconductors. The Fe/FeSe − x Te x bilayerwith a high T c of 17.7 K showed strong intrinsic pinningfrom correlated ab -planes, since the Se(Te)-Se(Te) inter-layer distance is almost identical to the out-of-plane co-herence length at low temperatures. ACKNOWLEDGMENTS
The authors thank J. Scheiter and E. Reich for helpwith FIB cut samples and TEM, and E. Barbara forhelp with the AFM observation. We are also grate-ful to M. K¨uhnel and U. Besold for their technical sup-port and S. F¨ahler for his RHEED software. Thiswork was partially supported by DFG under Project no.BE 1749/13 and HA 5934/3-1. We also acknowledge theEU (IRON-SEA and SUPERIRON) under Project no.FP7-283141 and FP7-283204. S. W. acknowledges sup-port by DFG under the Emmy-Noether program (Grantno. WU595/3-1). M. J. Wang, J. Y. Luo, T. W. Huang, H. H. Chang, T. K. Chen,F. C. Hsu, C. T. Wu, P. M. Wu, A. M. Chang, and M. K. Wu,Phys. Rev. Lett. , 117002 (2009). W. Si, Z. Lin, Q. Jie, W. Yin, J. Zhou, G. Gu, P. D. Johnson,and Q. Li, Appl. Phys. Lett. , 052504 (2009). E. Bellingeri, R. Buzio, A. Gerbi, D. Marr´e, S. Congiu, M. R.Cimberle, M. Tropeano, A. S. Siri, A. Palenzona, and C. Fer-deghini, Supercond. Sci. Technol. , 105007 (2009). P. Mele, K. Matsumoto, Y. Haruyama, M. Mukaida, Y. Yoshida, Y. Ichino, T. Kiss, and A. Ichinose, Supercond. Sci. Technol. ,052001 (2010). S. X. Huang, C. L. Chien, V. Thampy, and C. Broholm, Phys.Rev. Lett. , 217002 (2010). S. Agatsuma, T. Yamagishi, S. Takeda, and M. Naito, PhysicaC , 1468 (2010). E. Bellingeri, I. Pallecchi, R. Buzio, A. Gerbi, D. Marr´e, M. R.Cimberle, M. Tropeano, M. Putti, A. Palenzona, and C. Fer-deghini, Appl. Phys. Lett. , 102512 (2010). Y. Han, W. Y. Li, L. X. Cao, X. Y. Wang, B. Xu, B. R. Zhao,Y. Q. Guo, and J. L. Yang, Phys. Rev. Lett. , 017003 (2010). Y. Imai, T. Akiike, M. Hanawa, I. Tsukada, A. Ichinose,A. Maeda, T. Hikage, T. Kawaguchi, and H. Ikuta, AppliedPhys. Express , 043102 (2010). M. Hanawa, A. Ichinose, S. Komiya, I. Tsukada, T. Akiike,Y. Imai, T. Hikage, T. Kawaguchi, H. Ikuta, and A. Maeda,Jap. J. Appl. Phys. , 053101 (2011). I. Tsukada, M. Hanawa, T. Akiike, F. Nabeshima, Y. Imai,A. Ichinose, S. Komiya, T. Hikage, T. Kawaguchi, H. Ikuta, andA. Maeda, Appl. Phys. Express , 053101 (2011). T. Thersleff, K. Iida, S. Haindl, M. Kidszun, D. Pohl, A. Hart-mann, F. Kurth, J. H¨anisch, R. H¨uhne, B. Rellinghaus,L. Schultz, and B. Holzapfel, Appl. Phys. Lett. , 022506(2010). K. Iida, S. Haindl, T. Thersleff, J. H¨anisch, F. Kurth, M. Kid-szun, R. H¨uhne, I. M¨onch, L. Schultz, B. Holzapfel, andR. Heller, Appl. Phys. Lett. , 172507 (2010). M. Schulze, M. Hacisalihoglu, C. Blum, C. Hess, M. Kumar,A. Wolter, S. Wurmehl, and B. B¨uchner, “Systematic inves-tigations of the ternary system Fe-Te-Se (abstract, TT 10.28),”(2011), 75th Annual Meetings of the DPG and DFG Spring Meet-ing, Dresden, Germany. E. J. Kramer, J. Appl. Phys. , 1360 (1973). M. Kidszun, S. Haindl, T. Thersleff, J. H¨anisch, A. Kauffmann,K. Iida, J. Freudenberger, L. Schultz, and B. Holzapfel, Phys.Rev. Lett. , 137001 (2011). C. Uher, J. L. Cohn, and I. K. Schuller, Phys. Rev. B , 4906(1986). M. Eisterer, R. Raunicher, H. W. Weber, E. Bellingeri, M. R.Cimberle, I. Pallecchi, M. Putti, and C. Ferdeghini, Supercond.Sci. Technol.24