Table Of ContentThe influence of the buffer layer architecture on transport properties for
BaFe Co As films on technical substrates.
1.8 0.2 2
S. Trommler,1,a) R. Hühne,1 J. Hänisch,1 E. Reich,1 K. Iida,1 S. Haindl,1 V. Matias,2 L. Schultz,1,a) and B.
Holzapfel1,b)
1)IFW Dresden, P.O. Box 270116, 01171 Dresden, Germany, electronic mail: [email protected]
2)Superconductivity Technology Center, Los Alamos National Laboratory, Los Alamos,
New Mexico 87545; now at iBeam Materials, Santa Fe, New Mexico, electronic mail: [email protected]
A low and almost temperature independent resistance in the normal state and an anomalous peak effect
within the normal-superconducting transition have been observed in BaFe Co As /Fe bilayers, prepared
1.8 0.2 2
on IBAD-MgO/Y O buffered technical substrates. A resistor network array sufficiently reproduces this
2 3
2 effect, assuming an increase of the electrical conductance between tape and film with decreasing buffer layer
1
thickness. Based on this model we evaluated the influence of this effect on the critical current density and
0
successfully reconstructed the superconducting transition of the bilayer.
2
n
a The development of the coated conductors techno- commonly used to prevent oxide superconductors from
J
logy proved high temperature superconductors to be- deterioratingthesuperconductingpropertiesduetoare-
3 come commercially attractive in high current applica- duced diffusion of transition metal ions from the tape
2
tions. Basically two approaches have been focused on: into the film8.
rolling assisted biaxially textured substrates (RABiTs)
]
n and ion beam assisted deposition (IBAD) of textured
o buffer layers on polycrystalline metal tapes1,2. Re-
c
cently, the growth of Fe-based superconductors such as
-
r BaFe Co As (Ba-122) and FeSe Te on IBAD-
2−x x 2 0.5 0.5
p
MgO/Y O /Hastelloy has been demonstrated. Their
u 2 3
high critical current densities at liquid He temperature
s
. as well as their high upper critical magnetic fields make
t
a them suitable for high-current applications3–5. In this
m letter we investigate the normal-superconducting trans-
- ition for Ba-122 prepared on IBAD-MgO/Y2O3 buffered
d
Hastelloy tapes with varying Y O thickness. We will
n 2 3
point out that the observed peak in the resistive trans-
o
ition can be modeled and quantitatively explained us-
c
[ ing a simple resistor network attributing good electrical
conductance between Ba-122/Fe bilayer and tape in the
1
case of a thinner buffer stack. Finally, we show that this
v
model enables an accurate evaluation of important tech-
5
4 nical properties, such as critical temperature (Tc) and
4 critical current density (J ).
c
1 For comparison we used two different buffer architec-
2. tures on polycrystalline C-276 tape (Hastelloy). Tape
0 (A) was planarized by solution deposition planarization Figure1. (a)Layerarchitecturefortape(A)andcorrespond-
2 (SDP) process using multiple chemical solution coatings ingresistivesuperconductingtransition;(b)layerarchitecture
1 ofY O withatotalthicknessofapproximately1µm. To for tape (B) and corresponding superconducting transition.
2 3
v: obtain a biaxially textured buffer a MgO seed layer was The solid line represents a fit according the introduced res-
i preparedbyIBADatroomtemperaturecoveredwithho- istor network.
X
moepitaxial MgO of about 40 nm at 600 ◦C by electron
r beam sublimation6. Tape (B) utilizes electro-polished For both tapes, (A) and (B), a Ba-122/Fe bilayer was
a
Hastelloy, covered by a 10 nm thick amorphous electron- prepared in a pulsed laser deposition (PLD) setup with
beam-deposited Y O layer. On the 5 nm IBAD-MgO a base pressure of 10−9 mbar applying in-situ reflection
2 3
template160nmhomoepitaxialMgOwasdepositedwith high energy diffraction (RHEED) to monitor the film
electron beam sublimation7. The thicker Y O buffer is evolution3. 70 nm Ba-122 were deposited on 25 nm
2 3
Fe, which was shown to be beneficial for the textured
growth of Ba-1229. After cooling to room temperat-
ure the sample was covered with Au without breaking
a)DresdenUniversityofTechnology,DepartmentofPhysics,Insti- the vacuum to ensure low contact resistance. X-ray dif-
tuteforPhysicsofSolids,01062Dresden,Germany fraction proves the phase formation of Ba-122, whereas
b)TUBergakademieFreiberg,09596Freiberg,Germany pole figure measurements as well as RHEED verify the
2
epitaxial relationship (001)[100]Ba-122 (cid:107) (001)[110]Fe (cid:107) theunknownCocontentoftheFelayeritsresistivityρ
Fe
(001)[100]MgO3. Transport measurements were per- was estimated to a upper limit of 10 µΩcm. In general,
formedinfour-probegeometryusingaPhysicalProperty both, ρ and ρ , are temperature dependent functions,
n Fe
MeasurementSystem(PPMS;QuantumDesign). Forthe but were set constant in first-order approximation.
determination of J tracks were prepared by ion etching ThethicknessofFeinBa-122/Febilayershastoexceed
c
with 1 mm width and 1 mm length. the threshold for a closed layer, which is between 4 nm
Thenormal-superconductingtransitionofBa-122films and15nm,toensureacompletecoveragefilmfortheepi-
on the tapes (A), displayed in fig.1(a), have a typical taxialgrowthofBa-12210. Hence,theresultingthickness
shape comparable to films prepared on single crystals10. ratios (d) are in the order of 0.1 including an important
In contrast, films prepared on substrate type (B) (thin consequence, displayed in figure 2(b): The width of the
Y O layer) exhibit a remarkable shape of the transition superconducting transition of the Ba-122/Fe bilayer de-
2 3
as illustrated in fig.1(b). The resistance in the normal creases with increasing Fe layer thickness. Therefore, its
stateisthreeordersofmagnitudesmallerwithalmostno transition temperature is not identical to the single Ba-
temperaturedependence. Closeabovethesuperconduct- 122layerusingthesamecriterion. Ithastobeconsidered
ing transition the resistivity peaks, followed by a sharp thattheevaluatedT (B)fromtransportmeasurementsof
c
drop to zero. These features are common for films pre- bilayersdiffersfromthevaluesofthebaresuperconduct-
pared on tapes (B) independently of T and preparation ing compound, which is especially important regarding
c
conditions. Hence, they are a consequence of the tape application of multilayered film architectures.
architecture. Peak effects in the temperature depend- The description of the resistive transition for tape (B)
ent resistivity have also been observed in underdoped needsadditionalconsiderationstobetakenintoaccount.
BaFe Co As . However, they are accompanied by a A significant consequence of the varying buffer thick-
2−x x 2
reduced T and can not explain our observations11. ness is the influence on the electrical barrier properties
c
between superconductor and tape. In the case of tape
(B) an electrical connection can arise from defects in
tape which give rise to pinholes in the buffer. This pre-
sumptionissupportedbythefactthatincontrasttothe
used electro-polished tape comparable buffer stacks on
smoother mechanically polished tapes are completely in-
sulating. Additionally diffusion of metal ions from tape
and Fe buffer can increase the conductivity of the MgO
buffer. A thick Y O buffer suppresses the diffusion and
2 3
overgrowsthedefects. Inthefollowingwewillshowthat
achangeoftheoxidebufferlayerresistancebytwoorders
of magnitude gives rise to the observed resistance peak
at T for tape (B).
c
Figure 2. The simulated resistivity of the Ba-122/Fe bilayer AboveT thecurrentisshuntedbythetape,iftheBa-
c
(ρBL) decreases with increasing Fe layer thickness tFe to Ba- 122 film and metal tape are electrically good connected,
122 layer thickness tBa−122 ratio d (a). As consequence the astheresistivityofBa-122isoneorderofmagnitudelar-
superconducting transition width reduces (b) and the Tc,90 ger compared to Hastelloy11,13. Therefore, most of the
criterion shifts to lower temperatures.
bias current flows through the tape, which gives only a
small voltage drop between the voltage leads. Within
The reduction of the normal state resistivity of Ba-
the superconducting transition the fraction of current
122/Fe bilayers (ρ ) with increasing ratio in thickness
BL through the film increases with decreasing resistance of
between Fe (t ) and Ba-122 (t ) was reported
Fe Ba−122 the Ba-122 film resulting in an increasing voltage drop
previously10. This effect is demonstrated in fig.2(a), as-
between the voltage leads, which finally reaches zero
sumingasimpleparallelcircuitrepresentingtheresistiv-
when the film becomes fully superconducting. Taking
ityoftheFe(ρ )andtheBa-122layer(ρ ), where
Fe Ba−122 this into account, we split the sample geometry into
d = t /t . The superconducting transition can be
Fe Ba−122 threesegments,separatedbythevoltagecontacts,which
described as:
is schematically illustrated in fig.3(a). Based on these
assumptions, a resistor network was modeled in which
the conductance between tape and Ba-122/Fe bilayer is
ρ =ρ exp((−T/T )−q) (1)
Ba−122 n c given by the resistivity of the MgO/Y O buffer stack.
2 3
For a homogeneous sample with constant width and an
where ρ ≈ 1 mΩcm corresponds to the Ba-122 res-
n equidistantleadarrangementweendupintheexpression
istivity in the normal state and was taken from single
for the current I passing the voltage leads:
crystal data11. The parameter q determines the width
of the transition and is set to 25 in our case. The res-
istivity of pure Fe amounts to 0.01 µΩcm at 20 K and I =I 2Rb2 +4RbRt+Rt(RBL+Rt) (2)
increasesabout1µΩcmper1at%Codiffusion12. Dueto bias2R2 +(R +R )2+4R (R +R )
b BL t b BL t
3
where I is the bias current and R = ρ l/A is correspondingto tape(A). With decreasingbuffer thick-
bias BL BL
the resistance of Ba-122/Fe bilayer. A represents the ness by one order of magnitude also R /R decreases.
b n
film cross section. The segments have a width w = 10 Due to the reduced fraction of current flowing between
mmandalengthl=2mmcorrespondingtothedistance the voltage leads for R /R =1 the normal state resist-
b n
between the contacts. R and R denote the resistance ance is reduced, but no change in the transition shape
b t
of buffer and tape, respectively. In general R and R occurs. A further decrease to R /R = 0.1 corresponds
b t b n
are temperature dependent functions. However, we set to a good electrical connection between tape and super-
them constant in first-order approximation due to the conducting film equivalent to tape (B). In this case the
small temperature range which is treated. The result- normal state resistance is drastically reduced and a peak
ing voltage drop between the voltage leads is given by appears next to the normal-superconducting transition
U = R I = RI , where R is the resultant measured in agreement with the observed behavior. This calcula-
BL bias
resistance. TheresistivityofpolycrystallineHastelloyex- tion confirms that a change in the electrical connection
hibitsonlyasmalltemperaturedependencyandchanges between tape and Ba-122/Fe bilayer gives rise to the de-
in total 3.5% when cooled from room temperature to 10 scribed behavior.
K13. If the current is shunted in the normal state, the For R /R → 0 the normal state resistance converges
b n
transport properties are dominated by the tape, which to R and the peak in resistance vanishes as the resistor
t
explains the negligible temperature dependence of our network collapses to a simple parallel circuit. This case
data. With the resistivity of Hastelloy at 20 K, ρ =1.23 corresponds to a direct connection of the superconduct-
t
µΩm, and a tape thickness of 70 µm we obtain R ≈3.5 ing film with a conducting substrate and was observed
t
mΩ. The temperature dependent resistance of the Ba- previouslyforBa-122filmspreparedonsinglecrystalline
122/Fe bilayer was combined to R using a parallel cir- SrTiO substrates14. The substrate becomes conducting
BL 3
cuit and eq.1 as described previously. Its normal state duetothelossofoxygeninSrTiO athightemperatures
3
resistance R was set to 1 Ω, corresponding to typical in high vacuum, which results in a three orders of mag-
n
values of Ba-122/Fe bilayers, prepared on MgO single nitudelowernormalstateresistivityandareducedwidth
crystals (MgO-SC). of the superconducting transition compared to films pre-
pared on non-conducting oxide substrates. This estim-
ation suggests that the observed peak arises only in a
small range of the electrical barrier resistance between
Ba-122/Fe bilayer and tape. An expansion of the model
withadditionalresistorloopsforthesimulationofacon-
tinuousshuntingobbilayerandtapewithinthezonesA,
B and C does not affect the results above.
Figure 3. Schematic view of the four probe contact arrange- Figure 4. The calculated current which is shunted by the
ment(a)includingtheassumedresistornetworkmodel. Sim- tape depending on the critical current density, determined
ulation of the resistance, arising between the voltage leads at Ec = 1µ V/cm (a) and measured field dependent critical
for different barrier resistances (b) using eq. (2) and ensuing current density of the sample (inset). Reconstructed trans-
solutions R and R for R (c). The arrows indicate the itionoftheBa-122/Febilayerfrommeasurementdataoftape
+ − BL
resistive transition combining the two solution branches. (B)(fig.1(b)) using the resistor network in comparison to a
film prepared on a MgO single crystal (MgO-SC).
Considering the difference of the buffer thickness for
tape (A) and (B) the simulated resistive transition for Inthefollowingsectionweaddresstheaccurateestim-
decreasing R /R is presented in fig.3(b). For the cal- ationofthesuperconductingpropertiesoftheBa-122/Fe
b n
culation of the resistive transition we set T =23 K and bilayerfortape(B).Solvingthequadraticequationgiven
c
q =25ineq.(1). R /R =10correspondstoapoorelec- by eq.(2) results in two solutions for R denoted as R
b n BL +
tricalconnectionbetweentapeandsuperconductingfilm, and R , which have to be combined in order to obtain
−
4
the complete transition. This is illustrated in fig.3(c) for rier resistance between tape and Ba-122/Fe bilayer is re-
R /R = 0.1. The arrows indicate the resistive trans- duced. Asimpleresistornetworkarraydescribesallmain
b n
ition combining both solutions. features of the measured behavior. The error in the de-
For the reconstruction of the transition from experi- termination of the self field J is negligible for values
c
mental data reasonable resistance values for buffer and higher than 1 kA/cm2. The results indicate that a re-
tape have to be evaluated. In the normal state (T = 30 duced buffer layer thickness has no detrimental effect on
K) the resistance of the bilayer is larger than that of the superconducting properties of Ba-122. A thin elec-
tape and MgO/Y O barrier. Therefore, we can as- trical conducting barrier layer may even be beneficial as
2 3
sume R (cid:29) R ,R which gives R ∼ R(30 K) = 4 itprovidesaneffectiveprotectionofthesuperconducting
BL b t t
mΩ (fig.2(b)). This value is in good agreement with the filmincaseJ isexceeded,asthemajorityofthecurrent
c
estimated R using the resistivity of Hastelloy. will be shunted by the tape.
t
Since the combined resistive transition R has to be This work was partially supported by the Deutsche
BL
continuous, R and R have to be equal at their ex- Forschungsgesellschaft and SuperIron, founded by the
+ −
tremum which gives R ≈ 7.8 mΩ. The reconstruc- European Commission.
b
ted superconducting transition using these values is dis-
playedinfig.4(b)andcomparedtoaBa-122/Febilayeron
MgO-SCpreparedundercomparableconditions. TheT , REFERENCES
c
definedas50%ofthenormalresistivityat30K,waseval-
uated to 23 K which is slightly lower compared to films 1A. Goyal, M. Parans Paranthaman and U. Schoop, MRS Bull.
prepared on MgO-SC with T = 26 K. The origin of this 29(2004),552-561.
c
2Y.Iijima,K.Kakimoto,Y.Yamada,T.Izumi,T.SaitohandY.
deviationisnotuniquelyassignable. Both,areducedpre-
Shiohara,MRSBull.29(2004),564-571.
paration temperature while using metal substrates and
3K.Iida,J.Hänisch,S.Trommler,V.Matias,S.Haindl,F.Kurth,
additionaldoping ofBa-122bytransitionmetal ionswill I.L.delPozo,R.Hühne,M.Kidszun,J.Engelmann,L.Schultz,
affect T . The simulation of R is in very good agreement andB.Holzapfel,AppliedPhysicsExpress4(2011),013103.
c
withtheobservedtransitionoftape(B)usingeq.(2)with 4T.Katase,H.Hiramatsu,V.Matias,C.Sheehan,Y.Ishimaru,T.
Kamiya,K.TanabeandH.Hosono,App.Phys.Lett.98,(2011),
aBa-122/Fenormalstateresistanceof0.7Ωandthede-
242510.
termined values for Rb and Rt as displayed in fig.1(b). 5Q. Li, W. Si and I. K. Dimitrov, Rep. Prog. Phys. 74 (2011),
Finally, we etched a 1 mm wide and 1 mm long track 124510.
in section B (fig.3(a)) for the determination of J . We 6C.Sheehan,Y.Jung,T.Holesinger,D.M.Feldmann,C.Edney,
c
made the following assumption in order to estimate the J. F. Ihlefeld, P. G. Clem, and V. Matias, App. Phys. Lett. 98
(2011),071907.
influenceofthisgeometryonthefractionofcurrentflow-
7V.Matias,J.Hänisch,E.J.Rowley,andK.Güth,J.Mater.Res.
ing through the tape at the determined J of the track:
c 24(2009),125.
Currents for which the track enters the flux flow state 8V. Matias, E. J. Rowley, Y. Coulter, B. Maiorov,T. Holesinger,
arestilllossfreeintheremainingpartoftheBa-122sec- C. Yung, V. Glyantsev, B. Moeckly, SUST 23 (2010), 014018
tion A and C due their larger width. Therefore one can (5pp).
9T. Thersleff, K. Iida, S. Haindl, M. Kidszun, D. Pohl, A. Hart-
reduce the resistor network to a simple parallel circuit,
mann, F. Kurth, J. Hänisch, R. Hühne, B. Rellinghaus, L.
where the current through the track (Itrack) is defined as Schultz,andB.Holzapfel,App.Phys.Lett.97(2010),022506.
I =I +I andthefractionshuntedbythetape 10K. Iida, S. Haindl, T. Thersleff, J. Hänisch, F. Kurth, M. Kid-
bias track tape
is I = U /(R +2R ). The field criterion E = U /l szun,R.Hühne,I.Mönch,L.Schultz,B.Holzapfel,andR.Heller,
tape c t b c c
with the length of the track l = 1 mm was set to 10−6 Appl.Phys.Lett.97,(2010),172507.
11N.Ni,M.E.Tillman,J.-Q.Yan,A.Kracher,S.T.Hannahs,S.
V/cm for the determination of J . The resulting ratio of
c L.Bud’ko,andP.C.Canfield,Phys.Rev.B78(2008),214515.
Itrack/Ibias isgiveninfig.4(a),whereforJc higherthan1 12J. Bass, K.-H. Hellwege (ed.), J. L. Olsen (ed.)., Landolt-
kA/cm2 99% of the current passes the superconducting Börnstein - Group III Condensed Matter, Springer Materials,
track. For the self field J of the film, which is higher p.109&p.279,Fig.163-170,DOI:10.1007/b29240.
c 13J. Lu, E. S. Choi, and H. D. Zhou, J. App. Phys. 103, (2008),
than 70 kA/cm up to T = 18 K (fig.4(a), inset), this
064908
correction is negligible. 14K.Iida,J.Hänisch,R.Hühne,F.Kurth,M.Kidszun,S.Haindl,
In summary we have demonstrated that a peak in J. Werner, L. Schultz, and B. Holzapfel, App. Phys. Lett. 95,
the resistive transition appears when the electrical bar- (2009),192501.
