Analysis of coupling losses in multifilamentary untwisted BSCCOAg tapes through a.c. suscep
a r X i v :c o n d -m a t /0410548v 1 [c o n d -m a t .s u p r -c o n ] 21 O c t 2004
1
Analysis of coupling losses in
multi?lamentary untwisted BSCCO/Ag tapes through a.c.susceptibility measurements
D.Zola,F.G¨o m¨o ry,M.Polichetti,F.Strycek,J.Souc,P.Kov′a k and S.Pace
Abstract —Losses as function of the a.c.magnetic ?eld ampli-tude (B 0)were measured at 77K in untwisted BSCCO(2223)/Ag
tapes,at different frequencies,by measuring the imaginary part of the a.c.susceptibility.In particular,loss measurements were performed in the portions of the same tape,obtained by cutting it in pieces with different length,starting from around 12cm down to 1cm.The results show that the measured losses depend on the sample length but this observed behaviour is not always due to the coupling mechanism among the ?laments.In this work we discuss the observed experimental behaviour of different typology of tapes by analysing data comparing them with analytical models in order to fully characterize the tapes with regard to the coupling mechanism.
Index Terms —a.c.losses,BSCCO tapes,a.c.susceptivity
I.I NTRODUCTION
I
N A.C.electrical devices make by superconductors cables is necessary to reduce the a.c.losses arising by the hys-teretic,coupling and eddy mechanisms [1]–[3].Tapes or wires usually contains many superconducting ?laments because in this way the hysteretic losses are reduced.These ?laments in BSCCO tapes,are embedded in metallic matrix,usually in silver or silver alloy for improving the thermal stability and the mechanical properties of tapes.On the other hand,to cut down the coupling losses,it is essential to reduce the area of induced ?ux e.g.by twisting the ?laments or increasing the matrix resistivity or manufacturing arti?cial resistive barriers around the ?laments.Nevertheless,the efforts turned towards a reducing of losses in a tape can be compromised by the intergrowths and bridging which can decrease the effective resistivity of the matrix or electrically connect the supercon-ducting ?laments [4],[5].
Manuscript received 4October 2004.This work is partially supported by the European Commision project ENK6-CT-2002-80658ASTRA.
Danilo Zola is with Supermat,INFM Regional Laboratory and Department of Physics ”E.R.Caianiello”,University of Salerno,via S.Allende I-84081Baronissi (Salerno)Italy,e:mail zoldan@sa.infn.it,Tel.+39089965369,FAX :+39089965275.Fedor G¨o m¨o ry is with Institute of Electrical Engineering,Slovak Academy of Sciences,Dubravska Cesta 984239Bratislava,Slovakia.
Massimiliano Polichetti is with Supermat,INFM Regional Laboratory and Department of Physics ”E.R.Caianiello”,University of Salerno,Baronissi (Salerno)Italy.
Franticek Strycek is with Institute of Electrical Engineering,Slovak Academy of Sciences,Bratislava,Slovakia.
Jano Souc is with Institute of Electrical Engineering,Slovak Academy of Sciences,Bratislava,Slovakia.
Pavol Kovac is with Institute of Electrical Engineering,Slovak Academy of Sciences,Bratislava,Slovakia.
Sandro Pace is with Supermat,INFM Regional Laboratory and Department of Physics ”E.R.Caianiello”,University of Salerno Baronissi (Salerno)Italy.
The coupling loss per unit volume and per cycle (Q c )if the a.c.magnetic ?eld amplitude B 0?B p ,which is the full penetration ?eld of the tape,is given by [6],[7]:
Q c =
B 2
1+ω2τ2 (1)where τis so called time constant,χ0is a constant depending on geometry related to demagnetization factor and ωis the angular frequency.The Eq.1shows as the coupling loss depends on frequency and on B 0square whereas the hysteretic
loss depends on B 3
0[1],[8]and it has a very slight dependence on frequency.The different behaviour on frequency and on B 0can be employed to discriminate the different loss mechanism.The Equation 1has a maximum for ωτ=1which can be experimentally so estimated by ?nding the frequency (νm )where the maximum occurs in loss measurements performed at low B 0[9]–[11].
At the same time,loss are linked to the imaginary part of the ?rst harmonic of the a.c.susceptibility χ′′[3],[7],[12]according to:
Q =πχ′′χ0B 2
0/μ0(2)Both Q and τcan be also evaluated by a.c.susceptibility mea-surements also because χ0can be experimentally measured
[13]Moreover,in untwisted BSCCO tapes,τis given by:
τ=
?2μ0
2
not to vary as the length is changed(whereas the coupling loss depends strongly on the sample length),we want to characterize the quality of the tapes from bridging aspects. Finally,the measurement have been performed in a frequency range that the eddy loss have been estimated negligible.
II.E XPERIMENTAL
The a.c.losses were measured by using an a.c.susceptome-ter with a system of coils suitable for measurements on sample with length up to12cm An electromagnet produces an a.c. magnetic?eld with B0up to50mT,with a?eld homogeneity within1%on a8cm length and2%on12cm.The a.c.?eld induces a voltage in two racetrack-shaped coils:the pick-up coil,which is very close to the sample surface,while the null coil is1cm apart.Since the pick-up coil and the null coil are not perfectly identical,a variable compensation system is also used.The a.c.magnet and both the coils are placed in a reinforced plastic cryostat,so no eddy currents are induced in the cryostat walls.The system is cooled by liquid nitrogen and all the measurements have been performed at 77K.Measurements have been performed in the frequency range from1Hz to1000Hz in the?eld amplitude ranging from0.05mT to45mT.
A.C.susceptibility measurements have been performed on two bi-columnar tapes of around61mm.The?rst(named in the following”Ag tape”)was prepared with16?laments in pure silver matrix with a stack of8?laments for each column separated by about0.3mm of pure silver.The external sheath is also made with the same material.The geometry of the second tape(Ag/Mg tape)is similar to that of Ag tape,but the number of?laments is15and therefore there are8?laments in one column and7in the other.The metallic sheet between ?laments is a Ag/Mg(0.4%)alloy and the matrix which em-beds the whole?lamentary zone is a Ag/Mg(0.4%)/Ni(0.22%) alloy.Moreover,other a.c.susceptibility measurements have been performed on commercial tapes,manufactured by Aus-tralian Superconductor(AUS)and by Nordic Superconductor Technologies(NST),whose we have few technical data.The NST samples have65?laments embedded in unknown matrix (probably Ag-Mg alloy)and the cross section dimensions are 3.2mm×0.30mm.We do not know the?ll factor of the tapes neither the critical current.The Australian tapes have37?laments and a critical current of36-38A.The cross section is2.96mm×0.33mm,the?ll factor is unknown and the metallic matrix is probably in Ag/Mg alloy.
In all tape the geometrical factorχ0has been measured according the technique reported in Ref.[13].I measured value ofχ0are respectively8.9in Ag tape,8.8in Ag/Mg tape whereas it is3.6in NST tape and5.1in AUS tape.
III.F REQUENCY AND B0DEPENDENCE OF LOSSES IN
BICOLUMNAR TAPE
Losses have been measured as function of B0,in several pieces with different lengths,cut from our original Ag and Ag/Mg tapes.In the upper graph of the Fig.1,theχ′′(B0) is shown as measured on a Ag tape of61.7mm at different frequencies.For B0 10 10 10 c' ' Q ( J / m3 c y c l e ) B (T) Fig.1.B0dependence ofχ′′(upper graph)and of the loss(lower graph)in log-log scale as measured in Ag tape of61.7mm.The lines shown the expected dependence,for B0 χ′′has a maximum and it is around two time the full penetration?eld(B p)of tape),the susceptibility has a large frequency dependence.At low?eld amplitudes,χ′′is nearly constant and the slight dependence on B0is probably due to the hysteretic contribution of the superconducting columns. However,as shown in the lower graph of the same?gure,the loss densities have a quadratic slope as expected if the coupling mechanism leads through losses.At high?eld,the imaginary part and the Q take a behaviour similar to the hysteretic one due to a full coupling of the two superconducting columns[6], [10]. Losses have been investigated as function of the frequency at a?xed?eld amplitude much lower than B0,max.As reported in?gure2,the losses exhibit a maximum that shifts towards higher frequencies as the samples length decreases. The experimental data have been?tted by Q fit(ω)=α ωτ 1101001000 Q (J /m 3 c y c l e ) n (Hz) Fig.2.Frequency dependence of the losses of Ag tape (upper graph)and Ag/Mg tape (lower graph)as measured in samples of different lengths.The lines are the ?t obtained by using the (4)TABLE I V ALUES OF THE QUANTITIES τ,AND ρef f AS DETERMINED FROM THE EXPERIMENTAL DATA FOR ALL THE CONSIDERED SAMPLES . ?(mm)τ(ms)ρef f (μ?cm) Ag/Mg tape 61.8420.13230.98.90.15515.6 1.9 0.182 be understand knowing that high density of inter-growths is quite common tapes with matrix in Ag/Mg alloy [5]. IV.C OUPLING LOSSES IN COMMERCIAL TAPES In the previous section,we have shown as by mean a.c.susceptibility measurements,the bi-columnar tapes have been fully characterized,in particular evaluating the effective resis-tivity of the metallic matrix.On the same time,we have also veri?ed that the classical model on coupling losses works very well also if this is used to analyse the coupling mechanism in BSCCO/Ag tapes.We want to extend this analysis to com-mercial tapes which are the serious candidate for applications.In the ?gures 3and 4the losses measured on NST tapes and AUS tapes of different length are shown.Looking at the 10-410-310 -2 10 10 10 B 0 (T) 10 10 10 10 Q (J /m 3 c y c l e ) Fig.3.Losses as function of the magnetic ?eld amplitudes (in log-log scale)estimated on a NST multi?lamentary tape cut in pieces of different length,measured at different frequencies.In the inset,the χ′′(B 0)curves are shown,from which the losses have been estimated. Fig.3,we can observe in upper and lower graph that the losses have a cubic dependence at low ?eld whereas the slope becomes +1at high ?eld.This behaviour behaviour is observed on the sample 11.85cm long as in the sample of 1.6cm and equivalent results have been found at intermediate lengths.At the same time,the imaginary part of the a.c.susceptibility has not any frequency dependence.The analysis performed on these tapes lead us to conclude that in NST tape the coupling are completely suppressed for effect of a very strong bridging among the ?laments that lead to a behaviour very similar to a mono?lamentary tape with hysteretic loss only. In AUS tapes we can observe a more interesting behaviour.The losses have been measured on samples with length starting from 11.1cm down to 1.3.At low ?eld,the slope of the losses in log-log scale is close to 2.5as on sample 11.1cm long as on a piece of only 1.3cm.The value of the slope con?rms that the coupling mechanism competes with hysteretic one for every length.In the two inset of the ?gure,the imaginary part of the AUS tapes as measured in a piece of 11.1cm and 1.3cm show a some frequency dependence.Similar measurements,no shown,have been performed on a sample 5.5cm long and they have an equivalent behaviour.In particular in the two inset of Fig.4we can observe the low ?eld region of the a.c susceptibility measured at different frequencies.In the sample 11.1cm long,the χ′′decreases as the frequency increases and this could meaning that the maximum of the coupling losses is at lower frequency.As shown in previous section,cutting the sample this maximum shift at higher frequency.In the 5.5cm sample the χ′′seems to have a maximum between 7Hz and 21Hz whereas in the sample 1.3cm long this maximum 10 10 10 Q (J /m 3 c y c l e ) 10-410-310 -2 10 10 10 B 0 (T) Fig.4.Losses as function of the magnetic ?eld amplitudes (in log-log scale) estimated on a AUS multi?lamentary tape cut in pieces of different length,measured at different frequencies.In the inset,the χ′′(B 0)curves are shown,from which the losses have been estimated. seems to be at frequency around 189Hz with a saturation measured up to 567Hz.The frequency trend of AUS tape does not resemble exactly the quasi ideal behaviour of the bi-columnar tapes.On the other hand,we know that in pure metal the value of χ′′when ωτ=1is 0.38whereas it is,in general,smaller in a ?lamentary superconductor surrounded by a metal.For example in the bicolumnar tapes,where the coupling mechanism is dominant,at 0.1mT the maximum value of the susceptibility is around 0.15which is 3times larger than the value measured on AUS tapes at the same ?eld.All these considerations lead us to conclude that in AUS tapes a part of tape is strong bridged probably in the more densely packed region but an other part the bridge is not so strong and a coupling mechanism can arise.Nevertheless the large hysteretic terms is comparable with the coupling loss and we cannot fully characterize the samples.However,through this investigation we have acquired important informations on the a.c.behaviour of tapes potentially employed in the realization of applications investigating deeply the ?lamentary nature of the BSCCO tapes. V.C ONCLUSIONS In this work we have studied the a.c.coupling losses on two different sets of tapes,bicolumnar BSCCO tapes and commercial tapes.A.C.susceptibility and losses,measured as function of the magnetic ?eld amplitude and frequency,show that the coupling losses dominate in bi-columnar tapes over the hysteretic losses of the single ?laments.The exper-imental effective resistivity in Ag tape has a higher value in comparison with Ag/Mg tape which has a ρeff lower than expected.This results are explained with the presence of intergrowths in Ag/Mg tapes and this result suggests that in this kind of tapes,the enhancement of the effective matrix resistivity does not reduce automatically the coupling losses.In NST tapes we have found a very strong bridging which lead to suppress completely the coupling mechanism and the tape works like a mono?lamentary tape.Finally in AUS tapes the coupling mechanisms competes with a large hysteretic component,probably due to a strong bridging in a signi?catively part of the tape.Our work show as through a.c.susceptibility it is possible to investigate the quality of BSCCO/Ag multi?lamentary tapes. R EFERENCES [1]W.Carr,AC Loss and macroscopic theory of superconductors .New York:Gordon and Breach,1983.[2]S.Tak′a cs,“Ac losses in superconducting cables and their expected values in magnetic systems,”Superconductor Science and Technology ,vol.10,p.733,1997. [3]J.R.Clem,“Ac losses in type-ii superconductors,”in Magnetic Sus-ceptibilites of Superconductors and Other System ,E.by R.A.Hein T.L.Francavilla and D.H.Liebenberg,Eds.Plenum Press,New York and London,1991.[4]M.Dhall′e ,R.Passerini,E.Giannini,G.Witz,J.-Y .Genoud,Y .Huang, X.-D.Su,and R.Fl¨u ukiger,“The formation mechanism of bi,pb(2223)outgrwths in multi-?lamentary tapes,”IEEE Trans.Appl.Supercond ,vol.11,p.3557,2001. [5]Y .Bugoslavsky,Y .G.K.Perkins,P.Busceni,and A.Caplin,“Inhomo-geneity of current distribution in multi?lamentary biscco-2223tapes,”Superconducting Science Technology ,vol.14,p.245,2001. [6] A.M.Campbell,“A general treatment of losses in multi?lamentary superconductors,”Cryogenics ,vol.22,p.3,1982.[7] D.Zola, F.G¨o m¨o ry,M.Polichetti, F.Str′y ˇc ek, E.Seiler,I.Huˇs ek, P.Kov′a ˇc ,and S.Pace,“A study of coupling loss on bi-columnar bscco/ag tapes through ac susceptibility measuremets,”Susperconducting Science and Technology ,vol.17,p.501,2004. [8] C.Bean,“Magnetization of hard superconductors,”Review of Modern Physics ,vol.36,p.31,1964. [9]K.Kwasnitza,“ac losses of multi?lament superconductors reduced below the level of ?lament hysteresis losses due to ?eld screening by ?lament coupling currents,”Cryogenics ,vol.17,p.613,1977. [10]——,“scaling law for the ac losses of multi?lament superconductors,” Cryogenics ,vol.17,p.616,1977. [11]K.Kwasnitza and S.Clerc,“Ac losses of superconducting high-t c multi?lament bi-2223/ag sheated tapes in perpendicular magnetic ?elds,”Physica C ,vol.233,p.423,1994. [12]P.Fabbricatore,S.Farinon,S.Innocenti,and F.G¨o m¨o ry,“Magnetic ?ux shielding in superconducting strip arrays,”Physical Review B ,vol.61,p.6413,2000. [13]P.Fabbricatore,S.Farinon,F.G¨o m¨o ry,and S.Innocenti,“Ac losses in multi?lamentary high-tc tapes due to a perpendicular ac magnetic ?eld,”Superconducting Science and Technology ,vol.13,p.1327,2000. [14]M.Oomen,J.Rieger,M.Leghissa,and B.ten Haken,“Effective trans-verse resistivity in superconducting tapes with various ?lament shapes and structures,”Superconducting Science and Technology ,vol.13,p.1101,2000.