Progress in fabrication of second generation high temperature superconducting tape at Shanghai Superconductor Technology

The structural configuration of 2G-HTS tapes fabricated at SST is shown in figure 1, which is a typical ion beam assisted deposition (IBAD) + PLD layered structure. 10 mm wide, 50 μm thick C-276 Hastelloy tape is used as the substrate. Prior to film deposition, the Hastelloy tapes are first put through an electropolishing process to achieve an R_s of about 1 nm within a 1 μm by 1 μm region (measured by atomic force microscopy). A stack of buffer layers with the architecture of CeO₂/LaMnO₃/MgO/YO₃ is deposited. The layer composition and deposition techniques are indicated in figure 1. All of the oxide buffer layers are fabricated in a reel-to-reel manner with a tape travelling speed of about 70–400 m h⁻¹. The throughput of such a textured template is over 300 km/year for 10 mm wide tapes. Superconducting layers with various compositions are deposited by using a multi-plume multi-turn PLD system with a KrF excimer laser (LAMBDA 300 K). In our R&D process, optimization of the heating system for the large deposition area was performed in order to minimize temperature variation on the tape passing through the deposition zone. The details of this process and heating cylinder will be discussed in the following section. During deposition the oxygen partial pressure was set to be 300–500 mTorr (depending on the compositions of the targets), and the laser energy in the chamber was between 500–700 mJ with the repetition rate of the laser pulse being about 260–280 Hz. A superconducting layer with a thickness in the range from 1.5–2.5 μm can be fabricated by repeating the deposition two to four times. For power and magnet applications, three REBCO compositions were adopted, where GdBa₂Cu₃O₇ is used for 77 K, self-field applications, and the other two compositions are EuBCO +BaHfO₃ and YGdBCO phases (denoted as compositions A and B, respectively, hereafter), both being used for low temperature and high magnetic field applications. Lastly, a silver layer of 1–2 μm and a copper layer of 5–20 μm (depending on the user's requirement) are deposited by sputtering and electroplating techniques, respectively.

Zoom In Zoom Out Reset image size

The microstructure of the oxide buffer layers and superconducting layers are routinely characterized by optical microscopy, atomic force microscopy and x-ray diffraction analysis. A reel-to-reel accessory has been developed and installed in the x-ray diffractometer with an area detector, allowing it to scan and analyze the phase and texture of a short sample continuously. Due to the high reliability of the buffer layer stack fabrication process, routine quality control is carried out only on both ends of each tape. During IBAD-MgO layer deposition, an in situ reflected high energy electron diffraction (RHEED) detector is used to monitor the texture quality at the surface of the MgO layer. For short superconducting samples, I_c at 77 K, self-field was measured by the conventional DC four-probe method with the criterion of 1 μV cm⁻¹. Low temperature in-field I_c values were calculated based on the Bean model from the magnetization results, or obtained through direct conventional DC four-probe measurements. Longitudinal I_c distribution over long REBCO tape is measured by an MCorder^TM based on the inductive method [7]. For such inductive measurement, I_c values at both ends of the long tape were calibrated using four-probe measurement. Transmission electron microscopy (TEM) observations of the sample's cross section were performed using a Tecnai G2 F20 S-Twin operating at 200 kV.

The IBAD-MgO process is crucial for buffer layer fabrication. The surface texture of IBAD-MgO is very sensitive and plays an important role in the texture formation of the final cap layer. Figure 2 shows the RHEED patterns of the IBAD-MgO layer along a one kilometer-long tape, where the sharp patterns are obtained at the MgO surface. This result demonstrates the high stability of SST's IBAD process with the high tape traveling speed of 100 m h⁻¹. However, even with the sharp RHEED patterns observed, we found that CeO₂ (111) texture sometimes forms when directly deposited on the IBAD-MgO layers. With insertion of the RF-sputtered LaMnO₃ (LMO) layer prior to the CeO₂ layer, this issue has been resolved. Figure 3 shows the texture quality comparisons of the PLD-CeO₂ layer with and without the LMO layer among several batches of tapes produced. Evidently, the texture component in the CeO₂ layer with the LMO layer turns to single CeO₂ (002) orientation, i.e., the undesirable secondary texture component CeO₂ (111) disappeared. It is well known that CeO₂ (111) is fatal for further epitaxial growth of the REBCO layer and can result in a J_c reduction of more than 50% in our case. More interestingly, the out-of-plane texture also becomes sharper, while the in-plane texture remains unchanged. Considering that all the CeO₂ layers were deposited under the same conditions to the same thickness, the so-called 'self-epitaxy' effect [8] of the CeO₂ layer is principally the same as observed on the IBAD-MgO and on the sputtered LMO. So we deduced that the texture improvement is likely to be related to the stress releases or minor defect coverage by the LMO layer. The in-plane and out-of-plane texture degrees of the PLD-CeO₂ layers are routinely found to be 2°–4° and 2°, respectively, whilst the roughness is around 2 nm, as per atomic force microscope measurements.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Deposition temperature is one of the most crucial parameters for the PLD process. It is challenging to measure and analyze the practical deposition temperature when the tape is travelling, especially with high density plumes. Researchers from Bruker HTS and from Fujikura have proposed their solutions in their heating systems, i.e., quasi-equilibrium heater [9] and hot-wall heater [10, 11], respectively. In our PLD, a drum-type heating system is adopted in order to realize the large deposition zone with high travelling speeds. In such a system, the tape is wound on the steel cylinder turn by turn as described in our previous publication [12]. A schematic side view of the heater is shown in the top right corner of figure 4. The deposition zone is at the bottom part of the cylinder. The tape starts to be heated prior to touching, or when touching, the cylinder. As soon as the tape comes out from the right shield, REBCO begins to be deposited on the tape surface. The projection of the deposition zone can be approximated as a rectangle, i.e., the two sides of the rectangle are equivalent in length to the axial and azimuth directions of the cylinder, the area of which is related to the lane number and the arc of the cylinder exposed to the plumes. Herein, we studied the temperature variation along the arc of the cylinder, which is associated with the heating process and tape traveling speed. In our study, we noticed that the shields next to the heating cylinder have a significant effect on the tape temperature variation. Three factors—the geometry of the heating shield, the distance between the shield and the cylinder, as well as the tape travelling speed—have been taken into thorough consideration. In this study, we mainly optimize the geometry of the heating shield while maintaining high tape traveling speed. Figure 4 shows the temperature profiles measured on a tape with a high travelling speed of over 100 m h⁻¹ before and after optimization. Obviously, within the deposition zone the temperature variation is no greater than 10 °C, which is a remarkable improvement. It is believed that the improved temperature homogeneity is beneficial for the formation of thicker films and films with composite structure.

Zoom In Zoom Out Reset image size

Optimized deposition parameters are applied for the fabrication of long length REBCO tapes. By repeating the deposition (two to four times), film thicknesses in the range of 1–2.5 μm, depending on the I_c requirements of the users, can be achieved. We have selected GdBa₂Cu₃O₇ targets for 77 K, self-field applications. The longitudinal I_c distribution of a 300 meter piece of GdBa₂Cu₃O₇ tape is shown in figure 5. The average I_c is over 520 A cm⁻¹ width with standard deviation of less than 2%. The minimum and maximum I_c values are 504 A and 544 A, respectively. High I_c homogeneity can also be observed on a 4 mm wide tape after slitting. The local I_c drops (indicated by arrows in figure 5) observed are related to the defects on the metallic substrate. We have also investigated the target compositions (A and B) on the I_c(B) properties in the REBCO films, as shown in figure 6. Typically, we achieve I_c values of over 280 A/4 mm width at 30 K, 5 T and over 560 A/4 mm width at 4.2 K, 10 T (magnetic field being perpendicular to tape surface) in our production line. Anisotropy values of γ in the REBCO are around 2 (determined by the Blatter scaling approach [13–15]), which is much smaller than that for pristine YBCO (γ = 5 ∼ 7) [16, 17]. Through the XRD analysis, we did not observed nanorods along the c-axis in the REBCO films. Cross sectional TEM observation was performed, as shown in figure 6. It reveals a large amount of stacking faults instead of particles or column-like structures. This observation is further confirmed by the angular dependence of in-field critical current measurements shown in the inset of figure 7. A sharp J_c/${{J}_{{\rm{c}}}}^{sf}$ (θ) peak appears when the external magnetic field is parallel to the ab plane. This is attributed to both intrinsic pinning [18, 19] and extrinsic planar defects parallel to the ab-planes such as stacking faults [20, 21]. It is known that partial dislocations always form at the boundary between the stacking faults and REBCO matrix [22, 23]. Such defects could act as effective point pinning centers when the magnetic field is away from the ab-planes [24]. In particular, short length stacking faults increase density of the dislocations, leading to larger internal strains in our case. Therefore, the I_c anisotropy and in-field performance could be dramatically improved. One should notice that both films were deposited with the high growth rate (over 40 nm s⁻¹). Such microstructural characteristics might be strongly associated with the liquid phase during deposition, which needs to be studied further.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Besides superconductivity and I_c uniformity, lamination as well as the jointing techniques have also been extensively studied. Based on advanced lamination techniques [25], various lamination materials including stainless steel, brass, and copper can be used in order to fulfill different application requirements. Among them, one should notice that copper stabilized tape has the highest yield stress of over 1000 MPa at liquid nitrogen temperature, while the I_c retention is higher than 95% at a bending diameter of 5 mm (when the REBCO layer is facing inward). This data is comparable to other published data when 20–30 μm thick Hastelloy substrate is used in the 2G-HTS tapes [26]. The smaller critical bending diameter in our case is associated with the unique deposition conditions, as described previously. Namely, the REBCO layer in the tape is 'naturally' under compressive strain in the straight condition. Moreover, the 'dog-bone' defect of copper-plating has been overcome by shielding during the electroplating process. As shown in the inset of figure 8, the thickness nonuniformity of the copper layer is less than 10%. An overcurrent of 1200 A (in 100 ms) can be applied at 77 K on the tape with stainless steel lamination, while the room temperature resistance is also adjustable by optimizing the thickness combination of the stainless steel and copper stabilizer. Table 1 lists SST's existing jointing techniques and the characteristics, some of which have been published [27–29].

Zoom In Zoom Out Reset image size