Stress dependence of the critical currents in neutron 3 1 irradiated (RE)BCO coated conductors 0 2 n J Emhofer, M Eisterer, and H W Weber a J Atominstitut, Vienna University of Technology, Stadionallee 2, 1020 Vienna, Austria 8 1 E-mail: [email protected] ] n Abstract. The applicationof HTS coatedconductors infuture fusion or accelerator o magnetsiscurrentlyofincreasinginterest. HighLorentzforcesandthereforehighhoop c - stresses act on the conductors in large coils. The conductor is furthermore exposed to r p neutronradiationinfusionoracceleratormagnets. The expectedneutronfluence over u the desired lifetime of such magnets can be simulated by irradiation experiments in s a fission reactor. The coated conductors were characterized in the pristine state and . t 22 −2 a after irradiation to a fast neutron fluence of 1×10 m (ITER design fluence). The m sensitivity of the critical currents to applied stress was measured in liquid nitrogen. - Thecoldpartoftheset-upwaspositionedbetweenarotatablesplitcoilelectro-magnet d n forassessingtheIc-anisotropyupto1.4TundermaximumLorentzforceconfiguration. o The Ic-sensitivity to applied stress changed significantly in the GdBCO/IBAD- c conductors after irradiation, whereas nearly no change was observed in the [ YBCO/RABiTS-conductor. Furthermore, Ic and Tc were strongly reduced in the 1 GdBCO/IBAD-sample after irradiation. The angular dependence of Ic changed for v both samples in different ways after the irradiation, but no change in the angular 6 3 dependence was observed upon applying stress. 4 The highneutroncapturecross-sectionofGd andthe resultingstrongreductionof 4 . Tc seemtoberesponsibleforthe differentstressdependenceofIc inirradiatedGd-123 1 coated conductors. 0 3 1 : v i X r a Stress dependence of Ic in neutron irradiated coated conductors 2 1. Introduction Impressive progress in coated conductor development has been made in recent years[1]- [3]. The possibility of using liquid nitrogen as the coolant, the high mechanical stability[4]-[10] and the robustness of YBCO coated conductors against neutron irradiation[11][12] render YBCO coated conductors promising candidates for applications in future fusion or accelerator magnets. Tensile stress or strain affect the performance of the conductors reversibly up to the irreversible stress/strain limit(σirr / εirr). InthecaseofBi-2223conductorsthestress-dependence ofIc wasentirelyexplained by the pressure dependence of Tc[13]. However, since the pressure dependence of Tc is highly anisotropic in YBCO[14], the manufacturing method and in particular the alignment of the unit-cells strongly affect the stress dependence of Ic in YBCO. Recent studies [13][4] demonstrated that the reversible stress/strain-effects in YBCO/MOCVD and DyBCO/ISD conductors are also determined by the uni-axial pressure dependence of Tc. Furthermore, a strong field dependence of the Ic sensitivity on the applied tensile and/or compressive strain was found for YBCO/IBAD/MOCVD[4][5][8], YBCO/RABiTS/MOD[6][7] and YBCO/MOCVD/PLD/IBAD[10] samples. The change in Ic caused by neutron radiation depends on the superconducting material, the neutron energy, magnetic field and temperature. Fast neutrons (E>0.1MeV) generate collision cascades in cuprates by transferring sufficient energy to a primaryknock-on atom, which initiates further collisions [15][16]. These collisions lead to local melting of the lattice and the formation of spherical defects with a diameter of ≈ 6nm [15]. The size of these normal conducting impurities matches the size of a flux line core (2ξ ≈ 6nm at 77K) and is, therefore, optimal for flux pinning. a,b Furthermore, these defects are uncorrelated and randomly distributed. Fast-neutron- induced cascades have been shown to be responsible for flux pinning and intra-grain critical current enhancement in (RE)BCO [17][15]. Thermal and epi-thermal neutrons do not lead to extended defects in YBCO. Epi- thermal neutrons(in the keV range) may create point defects or clusters, if the recoil energy of the primary knock-on atom is high enough to displace at least one atom. These point defects disturb the regularity of the CuO2 planes,[18], which reduces the critical temperature. The reduction in Tc is rather weak in YBCO (about -2K for a 22 −2 fast neutron fluence Φ· t = 10 m [11], where Φ and t denote the fast neutron flux density and the irradiation time, respectively). Thermal neutrons(E < 0.5eV) do not affect the material by direct collisions, as the energy transferred by elastic collisions is below the displacement energy of a single atom. The neutron capture cross is negligible in YBCO. Nuclei with high neutron capture cross-sectionsσ(n,γ) should be avoided when the materials have to be exposed to neutron irradiation. Unfortunately, state-of-the- art (RE)BCO coated conductors contain gadolinium(GdBCO) or samarium(SmBCO) having high neutron capture cross-sections. The thermal neutron capture cross-sections 155 157 of the two stable isotopes Gd and Gd with a natural abundance of 14.8% and Stress dependence of Ic in neutron irradiated coated conductors 3 4 5 15.7%, respectively, are as high as 6.074 × 10 b and 2.537 × 10 b, respectively [19]. −28 2 (One barn, b, corresponds to 10 m .) After capturing a neutron, the excited nucleus emitsγ-raysandthecorrespondingrecoilenergyistransferredtothenucleus. Thisrecoil 156 158 is sufficiently high to displace the excited Gd and Gd nuclei thus creating a point defect. Neglecting the kinetic energy of the neutron, the recoil energy was calculated 156 158 to be 29eV and 34eV for the excited Gd and Gd nuclei[17]. Hence, in Gd or Sm-compounds the thermal and epi-thermal neutrons of the TRIGA reactor spectrum play a crucial role and should be removed for the simulation of a fusion spectrum by a fission reactor, since the expected fluence of low energy neutrons is very small at the magnet location of a fusion device. The results reported in this work refer to GdBCO/IBAD and YBCO/RABiTS conductors irradiated with the entire neutron energy spectrum of the TRIGA reactor[20]. 2. Samples Two different (RE)BCO coated conductors were used for this study. The SuperPower SCS4050 tape, which consists of a GdBa2Cu3O7−δ (GdBCO) superconducting layer on an IBAD(Ion Beam Assisted Deposition) template and the AMSC 344C(Amperium) tape consisting of an YBa2Cu3O7−δ (YBCO) superconductor on a RABiTS(Rolling Assisted Bi-axially Textured Substrate) template. The GdBCO layer in the GdBCO/IBAD sample was grown by MOCVD(Metal Organic Chemical Vapour Deposition) on a Hastelloy substrate with a thickness of 50µm. On top of the GdBCO-layer a 2µm silver layer was deposited and the entire sample was surrounded by 20µm copper. The individual samples were 50mm long, 4.04mm wide and 0.1mm thick. The REBCO layer in the YBCO/RABiTS contains yttrium and dysprosium with a nominal composition of Y:Dy:Ba:Cu of 1:0.5:2:3. The tape was grown by MOD(Metal Organic Deposition) onto a sputtered buffer stack. The cold worked and annealed (RABiTS) substrate of this tape was a 75µm thick Ni-5 at% W alloy. A ∼2µm silver layer was deposited on top of the YBCO layer and the entire tape laminated with a copper foil for better handling and stabilization. The samples were 50mm long, 4.4mm wide and 0.21mm thick. Note that different samples were used for the characterization in the pristine and the irradiated state, marked by #1 and #2 in the figures, respectively. Both samples were cut from the same spool and the self-field Ic at 77K as well as the homogeneity of the #2-samples were checked before the irradiation by four point measurements and magnetoscans, respectively. As the YBCO/RABiTS#2 was irreversibly damaged before reaching the irreversibility limits, a third YBCO/RABiTS sample denoted by #3 was used exclusively for a self-field Ic measurement in the unirradiated state. Furthermore, shorter reference samples of 26mm length from the same spools were irradiated in the same capsule for measurements of the irreversibility line. Stress dependence of Ic in neutron irradiated coated conductors 4 3. Experimental Details 3.1. Tensile stress set-up The cold part of our tensile stress insert(E in figure 1) was positioned between a rotatable split coil electro-magnet. Horizontal fields of up to 1.4T can be applied at different orientations(Θ). The angular range assessed in our experiments was about ◦ ◦ 220 , including both a,b-peaks, in angular steps of 1 around Hkc and Hka,b and ◦ 3 elsewhere. The cold part was immersed in liquid nitrogen during the measurements. Figure 1. Photograph of the tensile stress insert. (A)Stepper motor, (B)removable spring, (C)load cell, (D)connections, heating resistor, thermal switch, safety spring, etc., (E)cold part. The temperature variation of liquid nitrogen due to changes of ambient pressure or oxygen pick-up induces some scatter in the data. To reduce this error to an acceptable level, all data were discarded whenever the temperature at the sample position was not within 77.4±0.3K during the measurement. Tensile forces were set with a stepper motor(A in figure 1). A trapezoidal thread transformed the rotation to a translation along the pull rod axes. The force was transduced to the pull rod(B in figure 2) by a 1000kN spring(B in figure 1) and a Stress dependence of Ic in neutron irradiated coated conductors 5 ball bearing prevents the lower parts from rotating. A 2000kN load cell(C in figure 1) located between the ball bearing and the pull rod, measured the applied force. Figure 2. CAD-drawings of the lower cold part of the set up. The front side is illustrated in (a) whereas the back side, without the push rods and with an artificial cut-outinthecentre,isillustratedin(b). Wiresandcablesarenotdepicted. (A)Push rods, (B)pull rod, (C)removable carriage, (D)current clamps, (E)voltage contacts, (F)coated conductor, (G)Hall probe, (H)strain gauge, (I)reference strain gauge on a stress-free reference coated conductor. Figure 2 shows the front and back side of the cold part. The coated conductor(F in figure 2) is clamped between the upper and the lower part of the removable carriage(C in figure 2). The pull and the push rods also act as current bars and therefore brass was usedforthecoldpart. Theupper andlower partsoftheremovablecarriageareinsulated from each other by Teflon inserts. A Pt100-Class 1/5 sensor was used to monitor the temperature close to the sample. The voltage contacts(E in figure 2) were glued with silver paste, allowing small position rearrangements due to elongation. The average distance between the contacts was 5mm and Ic was evaluated using a 1µV/cm criterion by fitting the data by a power law(E = E (I/I )n, least squares method). c c A strain gauge was glued directly onto the coated conductor with epoxy resin(H in fig. 2), measuring the strain of the sample. A reference strain gauge was mounted close tothesample(Iinfigure2)onasmallstressfreecoatedconductorofthesametypeasthe measured sample in order to subtract the strain caused by cooling the sample to liquid nitrogen temperatures. Both strain gauges were connected to a tunable Wheatstone bridge. A Hall probe was mounted parallel to the tape surface in order to measure the orientation of the applied magnetic field, which was set by rotating the split coil with a stepper motor. Stress dependence of Ic in neutron irradiated coated conductors 6 The full Ic-anisotropy characterization at a fixed applied force was done during one measurement cycle. Such a cycle started by increasing the force stepwise from the relaxed state to the desired force. At each step, the self-field Ic was measured. Between each step, the stress was released and Ic was re-measured. The dependence of the self-field Ic on the stress/strain was obtained from these measurements. The anisotropy measurements started after reaching the desired set-force. Angular resolved Ic-measurements were performed at 0.1T, 0.2T, 0.4T, 0.6T, 1T and 1.4T. After the last measurement, the computer program started the warm-up process. Typically, anisotropy measurements were performed at nine different forces, leading to a net measurement time of typically around 400hours for one sample including periods of heating to room temperature. Note, that in the “relaxed” state a force of 10N(11MPa in YBCO/RABiTS and 25MPa in GdBCO/IBAD) was applied to the samples. 3.2. Neutron irradiation The samples were irradiated in the central irradiation facility of the TRIGA Mark II 22 −2 researchreactorinViennatoafastneutronfluenceofΦ·t=1×10 m (E > 0.1MeV). ◦ The samples were sealed in a quartz tube and the temperature did not exceed 70 C during the irradiation. The introduced defect structure in single-crystalline YBCO was already studied in detail [21] and further irradiation studies of coated conductors can be found in [12]. 3.3. Magnetoscan The homogeneity of the superconducting layer was checked by magnetoscan imaging before and after applying the maximum tensile stress. A detailed description of the set- up can be found in [22]. The local magnetic field detected slightly above the sample but below a permanent magnet is closely related to the local current density and, therefore, provides evidence for the local quality of the sample [23]. 4. Results and Discussion 4.1. Stress - strain dependence The irradiated YBCO/RABiTS sample (figure 3) was stressed up to the irreversible limits σ =410MPa and ε =0.52%. The yield strength of the substrate was irr irr reported to be 260MPa[24] at 76K which corresponds to a stress of 84MPa in the YBCO/RABiTS tape. The applied stress led to a distinct irreversible deformation at the strain gauge position after applying 346MPa (cf. figure 3). A reduction of the irreversible deformations was always observed after heating the sample to room temperature. The Young’s modulus(e) was evaluated by fitting the data between 20N(22MPa in YBCO/RABiTS and 50MPa in GdBCO/IBAD) and the maximum applied stress Stress dependence of Ic in neutron irradiated coated conductors 7 400 #YBCO/RABiTS 350 Φ.t = 1022 m-2 300 a) 250 87 MPa: e=120 GPa P 130 MPa: e=116 GPa σ (M 125000 127136 MMPPaa:: ee==110000 GGPPaa 260 MPa: e= 97 GPa 100 303 MPa: e= 96 GPa 346 MPa: e= 98 GPa 50 389 MPa: e=100 GPa 415 MPa: e=100 GPa 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 ε (%) Figure 3. Stress-strain dependence of the irradiated YBCO/RABiTS sample. Different symbols denote different maximum applied stresses. of the previous measurement loop (e.g. 173MPa for the 216MPa measurement). The values found for the irradiated YBCO/RABiTS sample are quoted in figure 3. No change of Young’s modulus was observed in the YBCO/RABiTS tape after irradiation, but a slight decrease by 24GPa was observed in the GdBCO/IBAD conductor. Since these results represent the first measurements ever on irradiated GdBCO/IBAD coated conductors and in view of the lack of sufficient statistics, further experiments will be needed for confirmation. The presently available average values of the Young’s modulus are listed in table 1. 4.2. Self-field Ic Self-field Ic measurements were performed together with the stress-strain measurements (except for the YBCO/RABiTS#3 sample). Figures 4 and 5 show the data up to the highest applied stress, where an irreversible reduction of Ic was observed for the first time. The critical currents are normalized by the mean value (before the samples were irreversibly damaged) of the critical currents in the relaxed state and at zero field. (A small force of 10N was applied also in the relaxed state in order to straighten the tape. Because of the different geometric cross sections of the tapes, this corresponds to a different pre-stress of 11MPa in the YBCO/MOD tape and of 25MPa in the GdBCO/IBAD tape.) The irreversible stress/strain limit was defined as the last value of applied stress/strain where Ic in the (following) relaxed state was above 95% of the initial value. 4.2.1. YBCO/RABiTS The stress dependence of Ic in the YBCO/RABiTS conductor did not change significantly after irradiation (figure 4). However, it seems that a small maximum appears at around 100MPa after irradiation. The pristine YBCO/RABiTS#1 sample was irreversibly damaged during the warm-up process after applying 303MPa, therefore no data exist for higher stress values. Below Stress dependence of Ic in neutron irradiated coated conductors 8 1.04 µH=0 T 0 a) 1.02 P M 1 1 1 T, 0.98 0 σ)/I(c 0.96 T, 0 0.94 YBCO/RABiTS#1 pristine I(c YBCO/RABiTS#3 pristine 0.92 YBCO/RABiTS#2 Φ.t = 1022 m-2 0.9 0 100 200 300 400 σ (MPa) Figure 4. Self-field Ic dependence on stress(YBCO/RABiTS). 303MPa(ε ≈0.33%), Ic was never reduced to below 98% of the original Ic. In the YBCO/RABiTS#3 sample, the irreversible stress limit was reached at 373MPa(ε ≈ 0.46%). Ic of the irradiated sample was reduced to 92%just below the irreversible stress limit of 410MPa(ε ≈ 0.53%). The relaxed Ic decreases after irradiation to 76% of Ic in the pristine sample. A Tc reduction of 1.8K was measured in the reference sample. Furthermore, B (77K) was irr only slightly enhanced at this fluence. 4.2.2. GdBCO/IBAD Figure 5 shows the Ic-stress dependence of the GdBCO/IBAD sample. 1.1 µH=0 T 0 a) 1 P M 5 2 0.9 T, 0 σ)/I(c 0.8 T, 0 I(c 0.7 GdBCO/IBAD#1 pristine #GdBCO/IBAD#2 Φ.t = 1022 m-2 0.6 0 200 400 600 800 σ (MPa) Figure 5. Self-field Ic dependence on stress(GdBCO/IBAD). After applying 740MPa(ε ≈0.55%), the critical current was permanently reduced in the pristine sample. Two major reductions occurred between 740MPa and 800MPa. After the first reduction, I in the relaxed state was reduced to ≈83% of its initial value. c After applying 790MPa, I inthe relaxed statewas reduced to ≈70%of the initial value. c Just below the irreversible stress limit of 740 MPa, Ic was reduced only by 5%. Note Stress dependence of Ic in neutron irradiated coated conductors 9 Table 1. Summary of the parameters obtained from the self-field measurements. (Tc from reference samples) Sample σirr ǫirr Tc Ic(0T,relaxed) n-value e pristine YBCO#1 >303MPa >0.33% 89.8K 88.4A 30.0 106GPa pristine YBCO#3 373MPa >0.46% 89.8K 93.6A 30.7 115GPa irradiated YBCO#2 410MPa 0.53% 88.0K 67.1A 25.8 103GPa pristine GdBCO#1 740MPa 0.55% 92.9K 142.7A 33.2 163GPa irradiated GdBCO#2 800MPa 0.65% 86.7K 16.52A 18.2 139GPa that measurements of other samples from the same spool indicated irreversible stress limits of up to 800 MPa. For the irradiated GdBCO/IBAD sample, the irreversible stress/strain limit was reached at 800MPa/ε=0.65%. Ic of the irradiated sample is more sensitive to applied stress than in the pristine sample. Close to the irreversible stress limit, Ic was reduced to 67% of the relaxed value. The relaxed self-field Ic was reduced by one order of magnitude after irradiation. Tc inthereferencesamplewasreducedby6.2Kandalsotheirreversibility fieldsignificantly decreased at this temperature. The lower Tc and the resulting higher T/Tc ratio of the irradiated tape at 77K seem to be responsible for the change of the Ic-sensitivity to applied stress. A higher Ic-sensitivity at temperatures closer to Tc was found for YBCO tapes in Ref. [10] and is expected from the equation for a single grain [25]: 1.5 T I (T,σ,b) = I (T = 0,σ = 0) 1− (1) c,sg c bσ +Tc(0)! where T (0) is the critical temperature at zero applied strain and b denotes the change c of Tc with applied pressure. The anisotropic pressure dependence of b was found to be b =-2.0±0.2K/GPa and b =1.9±0.2K/GPa with pressure along the a-axis and the a b b-axis, respectively [14]. For small σ equation 1 leads to first order to bT 2 Ic(σ)/Ic(0) = 1−1.5 σ +o(σ ). (2) Tc(Tc −T) The strain sensitivity thus increases with decreasing transition temperature, in particular, if Tc approaches T, as in the irradiated GdBCO/IBAD sample. Note that a pronounced curvature in Ic(σ) is observed experimentally instead of the predicted linear behavior. This is a consequence of the dependence of the critical currents on the applied electric field. A model with two grains oriented with their a-axis parallel and perpendicular to the tape direction was presented in [25], which describes the stress dependence of Ic in YBCO coated conductors qualitatively. However, since a real coated conductor is a percolative network of differently aligned grains, a quantitative model needs to be developed for a comparison with experimental data. Stress dependence of Ic in neutron irradiated coated conductors 10 4.3. Homogeneity Figure 6 shows magnetoscan images of the pristine GdBCO sample before and after the tensile stress measurements. The virgin conductor(figure 6a) is very homogeneous Figure 6. Magnetoscan images of the pristine GdBCO#1 sample before(a) and after(b) applying a tensile load of 320MPa. without significant defects. The conductor was irreversibly damaged after applying the maximum load of 740MPa. After removing the sample from the set-up, no destruction was found by optical inspection. The magnetoscan image(figure 6b) reveals the degree of damage. The superconducting layer was damaged over a long length and not only along single distinctive macroscopic cracks. The two main defects around x = −5mm and x = 10mm may correlate with the two significant reductions of Ic at ≈740MPa and ≈800MPa presented in figure 5. Note that the left and right ends, where the current clamps held the sample, are less damaged. Hence, the pressure exerted by the current terminals does not destroy the sample and a good current feed is ensured. 4.4. In-field Ic anisotropy The angular dependence of Ic at liquid nitrogen temperature changes completely after irradiation in both samples. In the YBCO/RABiTS conductor the strong additional pinning centres introduced by fast neutron irradiation lead to an enhancement of Ic between the two a,b peaks. Figure 7‡ shows the dependence of Ic in the pristine YBCO/RABiTS sample on the applied stress at µ H=1.4T. 0 The two maxima of Ic occur for fields parallel to the a,b-planes and the Ic- anisotropy(ratio of I (Hka,b)/I (Hkc)) is quite low. As expected from the self-field c c measurements, the applied stress hardly influences the critical current up to the maximum applied stress of 303MPa. For lower fields(not shown), the anisotropy is even smaller and no significant influence of applied stress was found either. ‡ Note that the color code in all following 3D-figures refers to a normalization of Ic by the (right) maximum near Hka,b (at 90◦and around 85◦in the YBCO/RABiTS and GdBCO/IBAD tape, ◦ respectively. 0 refers to the field parallel to the c-axis.)