Investigation of TiO barriers for their use in hybrid Josephson and tunneling junctions based on pnictide thin films
We tested oxidized titanium layers as barriers for hybrid Josephson junctions with high -products and for the preparation of junctions for tunneling spectroscopy. For that we firstly prepared junctions with conventional superconductor electrodes lead and niobium, respectively. By tuning the barrier thickness we were able to change the junctions’ behavior from a Josephson junction to tunnel-like behavior applicable for quasi-particle spectroscopy. Subsequently, we transferred the technology to junctions using Co-doped BaFeAs thin films prepared by pulsed laser deposition as base electrode and evaporated Pb as counter electrode. For barriers with a thickness of 1.5 nm we observe clear Josephson effects with 90 V at 4.2 K. These junctions behave SNS’-like and are dominated by Andreev reflection transport mechanism. For junctions with barrier thickness of 2.0 nm and higher no Josephson but SIS’- or SINS’-like behavior with a tunnel-like conductance spectrum was observed.
pacs:74.25.F-, 74.45.+c, 74.50 +r, 74.70.Xa, 85.25.-j
Josephson junctions are a potent tool for the investigation of iron pnictide superconductors. In our investigations on iron pnictide superconductors we prepared planar Josephson junctions using a sputtered insulating layer as barrier. Up to now different works for Josephson junctions with single crystals, Zhang et al. (2009a, b, 2012) for bi-crystal grain boundary junctions Lee et al. (2009); Katase et al. (2010, 2011); Hiramatsu et al. (2012); Schmidt et al. (2013) and edge-type junctions Döring et al. (2012a, b) as well as for planar junctions Schmidt et al. (2010) were published. The Josephson junctions presented therein used natural or engineered interfaces as well as normal metal layers as barrier, but up to now, no work with a deposited insulating barrier has been published. Additionally planar homo junctions with ferromagnetic barrier layers are in progress.Engelmann et al. (2013)
The behavior of a Josephson junction is most sensitive in the barrier properties e.g. its transparency. Our early experiments resulted in low resistance Josephson junctions with gold barriers showing an -product of about 18 V, Schmidt et al. (2010) which is much too low for possible applications like SQUIDs. Other works mentioned above and summarized in Ref. Seidel, 2011 showed -products of similar magnitude for different kinds of Josephson junctions with iron pnictides and various barriers. Only for junctions with a native barrier -products up to 300 V could be observed Zhang et al. (2009b).
One of the most famous phase-sensitive tests for the expected s-symmetry of iron pnictides from Parker and Mazin Parker and Mazin (2009) requires two distinct Josephson junctions with different transparencies. Thus, the transport of each junction will be dominated by the opposite charge carrier type so they will form a -SQUID. Hence, it is obvious to reach the idea by the combination of two junctions with a metal-like (high transparency) and insulating (low transparency) barrier, respectively.
Additionally, previous junctions with gold barriers showed non-ideal features like high excess currents, asymmetric - characteristics and non-constant background in microwave dependence of . Schmidt et al. (2010) By changing the barrier from normal conducting gold to insulating titanium oxide, an increase of the -products of our planar junctions is highly expected and a reduction of these non-ideal effects seems to be possible.
Another reason to introduce insulating barriers would be their use in quasi-particle spectroscopy on iron-pnictides. In low resistance junctions an additional series resistance from the electrodes can occur, when the measurement is done in the range of the transition temperature or with currents above of the film or electrode, respectively. Woods et al. (2004); Sheet, Mukhopadhyay, and Raychaudhuri (2004); Baltz et al. (2009); Chen, Huang, and Chien (2010) This spreading resistance can dramatically disturb the observed conductance spectra or at least impedes their normalization necessary for modeling within the BTK-theory. Blonder, Tinkham, and Klapwijk (1982) The effect of spreading resistance was observed in former junctions Schmidt et al. (2012) as well as in Point Contact Andreev Reflection (PCAR) studies on pnictides. Gonnelli et al. (2009a, b) Even if its possible to recalculate the real spectra with a semi-empiric model by characterization of the used electrodes, Döring et al. (2013) this means more expenditure and results only in a rough estimation. By using insulating barriers, it should be possible to produce junctions with much higher resistances. Thus spreading resistances from the electrodes should be negligible in these junctions.
For conventional superconductors titanium oxide barriers have been used to prepare Josephson or SIN-junctions. Celaschi, Geballe, and Hammond (1985); Otto, Tarasov, and Kuzmin (2007) Thus prior to the process on pnictide thin films, we used the barriers on junctions with two conventional superconductors lead and niobium, respectively, to test the barrier properties with well known materials.
Ii Junctions with conventional electrodes
For the preparation of these Nb\TiO\Pb junctions we use a SiO substrate with an e-beam evaporated niobium layer of 500 nm thickness. To prepare the junction area and the bonding pads of the Nb electrode we use a photo-lithography mask and subsequently ion beam etching (IBE) to remove material from the Nb surface of 30 nm depth. After this we deposit SiO by reactive sputtering with a thickness of 250 nm. The next step is the preparation of our counter electrode bonding pads with another photo-lithographic mask. For the pads we use a bi-layer system of titanium and gold with a thickness of 40 nm and 90 nm, respectively. The titanium is very adhesive on the sputtered SiO and gold acts very well as electrical and mechanical contact for future gold wire bonding. Both materials are sputtered in-situ at a pressure of 2 Pa in argon atmosphere and with a power of 50 W. After this we apply a photo-lithographic mask on the sample to deposit the barrier layer of oxidized Ti and the counter electrode layers of Pb and In. For the preparation of the barrier layer we clean up the Nb surface in the junction area from any oxides by removing 15 nm of the surface via IBE. Thereafter the titanium is deposited in-situ on the clean surface in Argon atmosphere at a pressure of 4 Pa and a sputtering power of 30 W at a deposition rate of 2 nm/min. Thus we have good control of the barrier thickness. We heat the sample in air up to 80 ˚C for 20 min to get a fast and controlled oxidation of the titanium. For the deposition of the counter electrode we use vacuum evaporation of lead with a layer thickness of 350 nm and followed by an indium protective layer with a thickness of 50 nm. A 3D-sketch and the corresponding cross-section of this junction is shown in fig. 1.
An - measurement for this junction type with a barrier thickness (TiO) = 2.5 nm is shown in fig. 2. One can see RSJ-like behavior with 50 A and 2.0 .
By increasing the barrier thickness it is possible to increase the resistance of the junction noticeably. In fig. 3 one can see a tunnel-like spectrum at K with a normal state conductivity of 75 10 or a resistance of 13.3 k, respectively. The spectrum, which was measured for a junction with (TiO) = 3.0 nm, shows peaks at meV and inner shoulders at meV, which are both marked by arrows in fig. 3. These values are a little bit lower than expected for ideal materials for this temperature. For all measured junctions and the subsequently analyzed spectra the fitted gap values were 1.04 meV 1.16 meV and 1.23 meV 1.36 meV, respectively, while the ideal values are 1.35 meV for lead and 1.45 meV for niobium. Thus a small degradation of the superconducting electrodes at the barrier interfaces can be assumed.Seidel and Richter (1980)
Iii Junctions with pnictide electrode
The preparation of the Ba-122\TiO\Pb junctions follows in principal our well established process for former junctions with gold barrier, which can be found in detail in Ref. Döring et al., 2012c. A first difference is the type of substrate we use for our pnictide thin film preparation. Instead of (La,Sr)(Al,Ta)O, we use CaF, which results in higher of the Co-doped BaFeAs (Ba-122) film. Kurth et al. (2013) Then we follow the processing steps from Ref. Döring et al., 2012c, covering the whole thin film with a 10 nm gold layer, pattering of base electrode via IBE, and forming the junction area by IBE and sputtering SiO frameworks. Now we use a separate photo-lithographic mask, which covers the whole sample except an area including each junction window plus some tolerance area around. Via IBE the covering gold thickness is reduced to nominally 1 nm. The reason for that is to reduce the influence of gold maximally, but not etch into the superconducting layer to avoid possible destruction of the superconductivity on the surface. Subsequently, we use a manipulator to transfer the sample from the IBE chamber into the sputtering chamber without breaking the vacuum. We sputter Ti layers with thicknesses between 1.0 and 3.0 nm. After the sputtering process, the sample is tempered and oxidized in atmosphere like described above. Finally, we continue with our usual step and evaporate a Pb film for the counter electrode followed by an In layer as protection. A sketch of this junction is shown in fig. 4.
In this work we will present results and show differences of samples with 1.5 nm and 2.0 nm of sputtered and oxidized titanium. Between these values there seems to be the transition from metal like to insulating behavior of the barrier. For barriers thinner than 1.0 nm we cannot exclude short cuts due to incomplete covering of the superconducting surface, while thicknesses above 3.5 nm result in resistances in the order of M.
iii.1 Josephson junctions
The first junction we present has a barrier of 1.5 nm thickness. As it can be seen in fig. 5, the temperature dependence of the resistance shows linear behavior with small saturation above the superconducting transition of Ba-122 at 26 K. This is typical for a metal-like barrier, thus we can assume a SNS’-junction has been formed by such thin barrier. Nevertheless, this junction shows Josephson effect which differs from our former ones with a gold barrier. Schmidt et al. (2010) Firstly, the --characteristic (see fig. 6) is not RSJ-like McCumber (1968); Lee (1971) but more similar to those Katase et al. observed for 16 grain boundary junctions. Katase et al. (2011) They show nearly linear behavior for . Opposite to the gold barrier junctions the - characteristic is nearly symmetric, in both branches equal values of 40 A and 2.2 can be estimated. This results in an of nearly 90 V which is higher than for the gold barrier junctions Schmidt et al. (2010) by a factor of 5 and at least 1.5 higher than for grain boundary junctions Hiramatsu et al. (2012), respectively. A detailed characterization of the prepared Josephson junctions and possible optimization of the barrier properties will be done in future work.
By changing the measurement setup from resistance versus current to conductance versus voltage one can obtain information about the order parameter of the used superconductors. In fig. 7 the conductance spectrum for the same junction as used for fig. 6 is shown. One can see a central peak for mV, with additional shoulders (see inset of fig. 7). Assuming a non-nodal order parameter this is a typical sign for transport dominated by Andreev-reflection and thus for a metallic barrier. Furthermore, some symmetric structures occur outside this central peak up to 20 mV in the conductance, which are presumable due to phonon interactions. Such features were also observed in the former PCAR studies Tortello et al. (2010, 2012) and are a footprint of the electron-boson spectral function . For the case of pnictide superconductors, they can be interpreted within a three-band Eliashberg model. Ummarino et al. (2011) For even higher voltages there is a parabolic decreasing background, which can be explained by Joule heating of the junction. Baltz et al. (2009)
iii.2 Tunneling junctions
By increasing the thickness of the TiO barrier from 1.5 nm the junction behavior changes dramatically. For a thickness of 2.0 nm some of the junctions on the same substrate still show Josephson effects while others do not. For even higher thicknesses we could not observe Josephson effect at all. In fig. 8 the temperature dependence of the junction resistance is shown for a junction with a 2.0 nm barrier without showing a Josephson effect. It can be seen that the resistance increases with decreasing temperature. These results infer that these junctions form either SIS’ or an SINS’ type ones. Also the superconducting transitions of the electrodes have only small (Ba-122) or quasi no (Pb) influence on the measured resistance, which is positive for the problem of spreading resistance.Döring et al. (2013)
In fig. 9 the corresponding --characteristic at K is shown and compared to the normal state resistance of 1.05 k. In contrast to fig. 6 it shows no Josephson effect but more tunnel-like behavior. This is supported by the conductance spectrum shown in fig. 10. It can be seen, that there is no central peak but a clear dip down to 80 percent of the normal state value. Additionally there are is only a small and noisy rise of the conductance beside the central dip. Such rise should be expected from energy dependence of the superconducting density of states (DoS). This behavior can be explained by two facts. Firstly, a high Dynes-parameter , which includes scattering processes into the BTK-model Dynes, Narayanamurti, and Garno (1978); Plecenik et al. (1994) and lowers features compared to pure ballistic junctions. Secondly, spin-polarization of the pnictide should be taken into account. In such an extended BTK-model, the DoS side peaks can be lowered or vanish completely for present spin-polarization for any type of junction barrier. Mazin, Golubov, and Nadgorny (2001); Auth et al. (2003)
The critical thickness of the barrier for junctions with a pnictide base electrode is about 1 nm lower than for junctions with the niobium base electrode. The possible reason for this is the lower roughness of the Ba-122 surface. While Ba-122 is an epitaxially grown thin film with nearly atomic flat surface, the niobium electrode is prepared from an e-beam evaporated layer on a Silicon wafer. Thus the niobium is amorphous with higher surface roughness than the Ba-122 thin film.
To summarize, we used TiO barriers firstly on junctions with conventional metal superconductors and showed that we were able to switch between Josephson effect and tunnel-like behavior by increasing the thickness of the barrier. Subsequently, we transferred the technology to junctions with iron pnictide base electrode and Pb counter electrode. For a double barrier with very thin Au and 1.5 nm TiO we could increase the resulted -product of SNS’-like Josephson junctions by at least a factor of 5 compared to former junctions with a pure 5 nm gold barrier. Additionally, we are able to prepare high resistant SIS’ or SINS’-like junctions for tunneling spectroscopy by further increasing the barrier thickness. With this first junctions Thus we were able to increase the properties and the possibilities of hybrid junctions with a pnictide electrode.
Acknowledgements.This work was partially supported by the DFG within SPP 1458 (project nos. SE664/15-2 and HA5934/3-1) and GRK 1621 and the European community under project IRON-SEA (project no. FP7-283141). Additionally, S.S. was funded by the Landesgraduiertenförderung Thüringen. We would like to thank R. Nawrodt for his useful comments.
- thanks: now at: Physikalisches Institut, Experimentalphysik II, Universität Tübingen, Auf der Morgenstelle 14, 72076 Tübingen, Germany
- thanks: now at: Institute for Technical Physics, Karlsruhe Institute of Technology, Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen, Germany
- X. Zhang, S. R. Saha, N. P. Butch, K. Kirshenbaum, J. Paglione, R. L. Greene, Y. Liu, L. Q. Yan, Y. S. Oh, K. H. Kim, and I. Takeuchi, Appl. Phys. Lett. 95, 062510 (2009a).
- X. Zhang, Y. S. Oh, Y. Liu, L. Yan, K. H. Kim, R. L. Greene, and I. Takeuchi, Phys. Rev. Lett. 102, 147002 (2009b).
- X. Zhang, B. Lee, S. Khim, K. H. Kim, R. L. Greene, and I. Takeuchi, Phys. Rev. B 85, 094521 (2012).
- S. Lee, J. Jiang, J. D. Weiss, C. M. Folkman, C. W. Bark, C. Tarantini, A. Xu, D. Abraimov, A. Polyanskii, C. T. Nelson, Y. Zhang, S. H. Baek, H. W. Jang, A. Yamamoto, F. Kametani, X. Q. Pan, E. E. Hellstrom, A. Gurevich, C. B. Eom, and D. C. Larbalestier, Appl. Phys. Lett. 95, 212505 (2009).
- T. Katase, Y. Ishimaru, A. Tsukamoto, H. Hiramatsu, T. Kamiya, K. Tanabe, and H. Hosono, Appl. Phys. Lett. 96, 142507 (2010).
- T. Katase, Y. Ishimaru, A. Tsukamoto, H. Hiramatsu, T. Kamiya, T. Keiichi, and H. Hosono, Nature Comm. 2, 409 (2011).
- H. Hiramatsu, T. Katase, Y. Ishimaru, A. Tsukamoto, T. Kamiya, K. Tanabe, and H. Hosono, Mat. Sci. Eng. B 177, 515 (2012).
- S. Schmidt, S. Döring, F. Schmidl, V. Tympel, S. Haindl, K. Iida, F. Kurth, B. Holzapfel, and P. Seidel, IEEE - Trans. Appl. Supercond. 23, 7300104 (2013).
- S. Döring, S. Schmidt, F. Schmidl, V. Tympel, S. Haindl, F. Kurth, K. Iida, I. Mönch, B. Holzapfel, and P. Seidel, Supercond. Sci. Technol. 25, 084020 (2012a).
- S. Döring, S. Schmidt, F. Schmidl, V. Tympel, S. Haindl, F. Kurth, K. Iida, I. Mönch, B. Holzapfel, and P. Seidel, Phys. Proc. 27, 296 (2012b).
- S. Schmidt, S. Döring, F. Schmidl, V. Grosse, P. Seidel, K. Iida, F. Kurth, H. S., I. Mönch, and B. Holzapfel, Appl. Phys. Lett. 97, 172504 (2010).
- J. Engelmann, K. Iida, F. Kurth, C. Behler, S. Oswald, R. Hühne, B. Holzapfel, L. Schultz, and S. Haindl, Physica C 494, 185 (2013), article in press.
- P. Seidel, Supercond. Sci. Technol. 24, 043001 (2011).
- D. Parker and I. I. Mazin, Phys. Rev. Lett. 102, 227007 (2009).
- G. Woods, R. Soulen, I. Mazin, B. Nadgorny, M. Osofsky, J. Sanders, H. Srikanth, W. Egelhoff, and R. Datla, Phys. Rev. B 70, 054416 (2004).
- G. Sheet, S. Mukhopadhyay, and P. Raychaudhuri, Phys. Rev. B 69, 134507 (2004).
- V. Baltz, A. D. Naylor, K. M. Seemann, W. Elder, S. Sheen, K. Westerholt, H. Zabel, G. Burnell, C. H. Marrows, and B. J. Hickey, J. Phys. - Condes. Matter 21, 095701 (2009).
- T. Y. Chen, S. X. Huang, and C. L. Chien, Phys. Rev. B 81, 214444 (2010).
- G. E. Blonder, M. Tinkham, and T. M. Klapwijk, Phys. Rev. B 25, 4515 (1982).
- S. Schmidt, S. Döring, V. Tympel, F. Schmidl, , S. Haindl, K. Iida, B. Holzapfel, and P. Seidel, Phys. Proc. 36, 82 (2012).
- R. S. Gonnelli, M. Tortello, D. Daghero, G. A. Ummarino, V. A. Stepanov, and J. S. Kim, Cent. Eur. J. Phys. 7, 251 (2009a).
- R. S. Gonnelli, D. Daghero, M. Tortello, G. A. Ummarino, V. A. Stepanov, R. K. Kremer, J. S. Kim, N. D. Zhigadlo, and J. Karpinski, Physica C 469, 512 (2009b).
- S. Döring, S. Schmidt, S. Gottwals, F. Schmidl, V. Tympel, I. Mönch, F. Kurth, K. Iida, B. Holzapfel, and P. Seidel, arXiv.org , 1309.1641 (2013).
- S. Celaschi, T. H. Geballe, and R. H. Hammond, J. Appl. Phys. 57, 1698 (1985).
- E. Otto, M. Tarasov, and L. Kuzmin, Supercond. Sci. Technol. 20, 865 (2007).
- P. Seidel and J. Richter, Phys. Status Solidi B 98, 189 (1980).
- S. Döring, S. Schmidt, F. Schmidl, V. Tympel, S. Haindl, F. Kurth, K. Iida, I. Mönch, B. Holzapfel, and P. Seidel, Physica C 478, 15 (2012c).
- F. Kurth, E. Reich, J. Haenisch, A. Ichinose, I. Tsukada, R. Huehne, S. Trommler, J. Engelmann, L. Schultz, B. Holzapfel, and K. Iida, Appl. Phys. Lett. 102, 142601 (2013).
- D. McCumber, J. Appl. Phys. 39, 3113 (1968).
- P. Lee, J. Appl. Phys. 42, 325 (1971).
- M. Tortello, D. Daghero, G. A. Ummarino, V. A. Stepanov, J. Jiang, J. D. Weiss, E. E. Hellstrom, and R. S. Gonnelli, Phys. Rev. Lett. 105, 237002 (2010).
- M. Tortello, D. Daghero, G. A. Ummarino, V. A. Stepanov, N. D. Zhigadlo, J. Karpinski, J. Jiang, and R. S. Gonnelli, J. Supercond. Nov. Magn. 25, 1297 (2012).
- G. Ummarino, D. Daghero, M. Tortello, and R. Gonnelli, J. Supercond. Nov. Magn. 24, 247 (2011).
- R. C. Dynes, V. Narayanamurti, and J. P. Garno, Phys. Rev. Lett. 41, 1509 (1978).
- A. Plecenik, M. Grajcar, S. Benacka, P. Seidel, and A. Pfuch, Phys. Rev. B 49, 10016 (1994).
- I. Mazin, A. Golubov, and B. Nadgorny, J. Appl. Phys. 89, 7676 (2001).
- N. Auth, G. Jakob, T. Block, and C. Felser, Phys. Rev. B 68, 024403 (2003).