High-temperature-tolerable superconducting Nb-alloy and its application to Pb- and Cd-free superconducting joints between NbTi and Nb3Sn wires

For more than 30 years, Pb–Bi alloy and Wood's metal (50% Bi, 26.7% Pb, 13.3% Sn, and 10% Cd) have been used as representative superconducting solder intermedia to establish superconducting joints between NbTi and Nb3Sn wires in high-field nuclear magnetic resonance magnet systems. However, the use of Pb and Cd has been severely restricted by environmental regulations, such as the Restriction of Hazardous Substances Directive. Herein, a novel method of forming a superconducting joint between NbTi and Nb3Sn wires without Pb and Cd has been proposed. This approach is based on metallurgical bonding processes using a superconducting Nb-alloy intermedium, whose fine microstructure is maintained even after exposure to temperatures higher than 650 °C. Further, fine crystal defects become sources of magnetic flux pinning centers. Among transition elements close to Nb, Hf is considered the most suitable additive for realizing high-temperature-tolerable (HTT) superconducting Nb-alloy intermedia. Utilizing the HTT characteristic of Nb–Hf, a superconducting joint between Nb3Sn filaments and one end of the Nb–Hf alloy core was created by forming a superconducting Nb3Sn layer at the interface through a chemical reaction. The other end of the Nb–Hf alloy core was cold-pressed with NbTi filaments, to connect their active new surfaces to each other in order to create a superconducting joint. Ultimately, a superconducting joint between NbTi and Nb3Sn wires was realized with a high critical magnetic field (Bc2) of more than 1 T. The formation of the superconducting joint was confirmed by current decay measurements. This method of forming a superconducting joint is promising for application in environmentally friendly nuclear magnetic resonance magnet systems.


GRAPHICAL ABSTRACT Introduction
Nuclear magnetic resonance (NMR) spectroscopy is an essential analytical tool in biology, chemistry, and materials science and requires operation in persistent current mode for temporal field stability and a high signal-to-noise ratio. To date, Pb-Bi and Wood's metal (50% Bi, 26.7% Pb, 13.3% Sn, and 10% Cd) have served as superconducting intermedia to join separated superconducting wires to each other in NMR magnets [1,2]. Pb-Bi and Wood's metal have relatively high critical magnetic fields (B c2 ) of more than 1.5 [3][4][5] and 1.0 T [6] at the liquid helium temperature, respectively. These superconducting solders are now indispensable materials, especially for achieving superconducting joints between NbTi and Nb 3 Sn wires in high-field NMR magnet systems [1,2].
However, Pb and Cd are toxic. Their use has been severely restricted by environmental regulations, such as the restriction of hazardous substances (RoHS) directive. The utilization of these materials is permitted only after an approval of exemption, and the preparation for this procedure is costly. Furthermore, the use of these materials may be entirely prohibited in future. Therefore, there is an urgent need to develop new methods for realizing Pb-and Cd-free superconducting joints in industry.
Two possible solutions are considered to overcome the aforementioned challenges. The first is the development of new Pb-and Cd-free superconducting solders. Based on this approach, InSn-based alloy was developed. [7][8][9]. Although this material was demonstrated to be a promising candidate, it only showed a maximum B c2 of approximately 0.2 T. Moreover, to date, there has been no report on the successful formation of a superconducting joint between NbTi and Nb 3 Sn wires using this material.
The second solution is the utilization of a metallurgical reaction process for bonding without a solder. However, this approach remains largely unexplored. Metallurgical processing without a solder for establishing superconducting joints between NbTi and Nb 3 Sn wires becomes difficult to achieve due to two main reasons. The first is the mechanical problem of Nb 3 Sn. NbTi is a body-centered cubic alloy in which Ti dissolves in Nb. Solid solution NbTi is ductile, which enables superconducting joints to be established by connecting their active new surfaces to each other through mechanical pressing [10][11][12]. In contrast, Nb 3 Sn is an intermetallic compound with an A15-type lattice structure [13,14], which is so brittle that mechanical pressing cannot be applied.
The second reason is the rapid deterioration of the superconducting properties of NbTi after exposure to high temperatures. Figure 1 shows critical current (I c )-B characteristics before and after heat treatment at 100 h/685 8C, where the wire employed is standard reference material (SRM) 1457 of the National Bureau of Standards using Nb-46.5wt%Ti (Nb-62.8%Ti). A superconducting joint between NbTi and Nb 3 Sn can be achieved via a chemical reaction process. In fact, we confirmed that the Nb 3 Sn layer can be formed at the interface as a superconducting pathway between the NbTi and Nb 3 Sn filaments through a Sn diffusion heat treatment process. However, the transport properties of the joint sample were unsatisfactory because of the degradation of the NbTi properties. Thus, there was no alternative to incorporating superconducting solders that turn into liquid at low temperatures, between the NbTi and the brittle Nb 3 Sn superconducting filaments without damage.
In view of this, a novel method utilizing a Nbbased ductile superconducting alloy as an intermedium was introduced in this study. The I c of this superconducting alloy does not deteriorate even after exposure to the Nb 3 Sn layer formation temperature, which is typically higher than 650°C. Utilizing this material, superconducting joints were established with NbTi through mechanical pressing and with Nb 3 Sn through a chemical reaction process to form a Nb 3 Sn superconducting layer at the interface; ultimately, a superconducting joint was realized between the NbTi and Nb 3 Sn wires.
The Nb alloy follows the Ginzburg-Landau-Abrikosov-Gorkov (GLAG) theory [15,16]. This theory has motivated many microstructural studies of Nb alloys related to their superconducting properties, dating back to the 1960s [17][18][19][20][21][22]. The GLAG theory predicts the upper critical field (B c2 ) at 0 K to be given by B c2 (0) = 3.11 9 10 3 q n cT c , where q n , c, and T c are the normal state resistivity, electronic specific heat coefficient, and critical temperature, respectively [15,16]. The severe deformation of the Nb-alloy solid solution phase leads to the development of a dense dislocation cell structure in the polycrystalline Nb alloy grain morphology following the formation of a preferred orientation or texture. This fine nanoscale microstructure increases q n , thereby increasing B c2 . In the case of NbTi, B c2 reaches 11 T from a bulk value of approximately 1.5 T. The fine microstructure also contributes to the improvement of the flux pinning property, which is essentially related to the critical current density (J c ) [23][24][25]. Furthermore, in the Nb-Ti system, precipitation of dense nanoscale a-Ti ribbons can also provide a large additional pinning force [26][27][28][29]. However, the fine microstructure becomes coarse after heat treatment at temperatures higher than 650°C.
Thus, it can be inferred that the key to developing such high-temperature-tolerable (HTT) superconducting Nb-alloys is to create a nanoscale fine microstructure in the alloy and maintain the microstructure even at high temperatures. Recently, Balachandran et al. [30] reported an interesting finding in their Nb 3 Sn work that the fine microstructure of the unreacted Nb-Ta-Hf core was maintained after Nb 3 Sn layer formation annealing. Following this, a group of researchers, which included the author of this paper, confirmed that Nb with a small addition of Hf has a high recovery temperature of more than 700°C through a Nb 3 Sn layer formation study [31]. This Nb alloy maintains a fine microstructure even after annealing at 800°C for 3 h. Figure 2 shows an electron backscatter diffraction (EBSD) inverse pole figure (IPF) map and image quality (IQ) map of the microstructure of Nb-4 at%Ta-1 at%Hf bulk tape with severe deformation, before and after annealing. Inspired by these findings, the author of this paper considered this material to be a promising candidate for HTT Nb alloys.
In this work, we explored promising HTT Nb alloys that can be used for establishing superconducting joints between NbTi and Nb 3 Sn wires, and successfully developed a Pb-and Cd-free superconducting joint between NbTi and Nb 3 Sn wires using the HTT Nb alloy. The realization of the superconducting joint was confirmed by ultra-low resistance measurements using a current decay measurement system [32].

Materials
HTT superconducting Nb-alloys were synthesized with transition metals of Groups 4B-6B, which comprise a ductile body-center-cubic (BCC) superconducting Nb alloy. In particular, Ti and Zr are known to be effective additive elements to improve the superconductivity of Nb substantially after deformation [17,20]. Bulk alloys consisting of two or three elements were first formed by arc-melting. The ascast bulks were cold-drawn by grooved-rolling. Subsequently, the bulks were re-arc-melted for homogenization.

Test wires to assess the superconducting properties of Nb alloys
A fine nanoscale microstructure, which is important for improving the superconducting properties, is created in the alloy by fabricating single-core composite thin tapes. The superconducting properties such as B c2 and I c of the tape wires are checked before and after heat treatment at 685°C for 100 h, which is a typical Nb 3 Sn layer formation condition. The NbTi SRM wire was used as is in a round wire shape. The arc-melted Nb-alloy bar was swaged down to [5.5 mm and annealed at 900°C for 5 h in a vacuum. The bar was inserted into a Cu tube with outer and inner diameters of 8 and 6 mm, respectively, and cold-drawn to a wire [1-1.34 mm. Then, single-core composite thin tapes were fabricated by flat-rolling the single-core Cu/Nb-alloy wire: the tape had a thickness of 0.16-0.33 mm and a width of 2.5-4.2 mm. The rolling reduction ratio was approximately 400-550%.
I c was measured at 4.2 K using a standard fourprobe method with a tap distance of 1 cm and defined with a criterion of 1 lV/cm. The magnetic field was applied perpendicular to the tape axis and parallel to the tape surface. B c2 was determined as a magnetic field at which I c becomes almost zero. The n-values (V µ I n , where V and I are the transition voltage and transport current, respectively) were obtained by fitting the V-I curve in the range of approximately 0.1-1.0 lV. Cryogenic measurements were performed at the Tsukuba Magnet Laboratory, National Institute for Materials Science (NIMS), Japan. Figure 3 shows a schematic illustration of the developed joining process and a fabricated joint sample. The process basically consists of two steps: in step 1, the Nb 3 Sn filaments and HTT Nb alloy are connected via a metallurgical chemical reaction, and in step 2, the HTT Nb alloy and NbTi filaments are joined via mechanical bonding.

Joining process
In step 1, at first, one end of the Nb 3 Sn precursor wire, which is to serve as one end of the Nb 3 Sn coil, is chemically etched via HNO 3 to remove the outer Cu sheath. Subsequently, the diffusion barrier inside the Nb 3 Sn precursor wire is mechanically polished to expose the Nb filaments. Typical practical Nb 3 Sn precursor wires have Cu stabilizers and diffusion reaction barriers of Nb or Ta around the multifilamentary region consisting of Nb filaments and a Cu-Sn matrix. The wire tested in this study was a typical internal-Sn Nb 3 Sn wire with a diameter of 0.8 mm, consisting of 1045 Nb filaments, which was fabricated in our laboratory. A detailed description of the wire configuration can be found in the literature [33].
Second, one end of the HTT Nb alloy is chemically etched via HNO 3 to remove the outer Cu sheath. Then, the end of the Nb 3 Sn precursor wire is wrapped spirally by the end of the HTT Nb-alloy tape, followed by pressing it with a supporting metal tube to connect the active new surfaces of the filaments and the tape core. In this study, a Nb tube with outer and inner diameters of 6 and 5 mm, respectively, and a length of 15 mm was used. The pressed length of all the joint parts was almost 1 cm, and the pressure was set approximately 120 kN.
Finally, the Nb 3 Sn layer formation heat treatment is performed at 685°C for 100 h in a vacuum. In the final stage, Sn diffuses from the Cu-Sn phase into not only the Nb filaments of the Nb 3 Sn precursor wire, but also the surface of the HTT Nb-alloy core and the interface between the Nb filaments and the HTT Nb alloy core, where it subsequently forms the Nb 3 Sn superconducting layer.
In step 2, at first, one end of the NbTi wire, which is to serve as one end of the NbTi coil, is chemically etched to remove the Cu matrix and expose the NbTi filaments. In typical NbTi wires, NbTi filaments are embedded into the Cu matrix and twisted. The exposed bundle of NbTi filaments is untwisted back before being connected with the HTT Nb-alloy by cold-pressing to avoid filament crossing as much as possible during pressing. The Cu sheath of the other end of the HTT Nb-alloy tape is chemically etched as well. The end of the NbTi wire is wrapped spirally by the end of the HTT Nb-alloy tape and pressed with a supporting metal tube to interconnect the active new surfaces of the filaments and tape core. In this study, a Nb tube with outer and inner diameters of 6 and 5 mm, respectively, and a length of 15 mm was used. The pressed length of all the joint parts was almost 1 cm, and the pressure was approximately 120 kN.
Nb has a fairly high O affinity, easily forming a stable thin oxide layer on the NbTi filaments and HTT Nb alloy surface. The existence of an oxide layer suppresses the connection area between new surfaces. Such poor physical connectivity could be improved through additional annealing at a moderate temperature of approximately 450°C. This additional annealing does not affect the properties of the Nb 3 Sn wire and the Nb 3 Sn layer formed at the joint interface, because this temperature range is much lower than the Nb 3 Sn layer formation temperature.

Microstructural and microchemical analysis
The microstructures were observed using a fieldemission scanning electron microscope (FESEM). Compositional analysis was performed by energy-  dispersive X-ray spectroscopy (EDS). Polished samples were prepared with a standard micrographic technique, where a sample was embedded in a conductive phenolic resin and finely polished with a 0.05 lm sol-gel alumina suspension. The grain orientation maps of the microstructure were measured by EBSD analysis.

Current decay measurement
In general, the resistance of a superconducting joint is lower than 10 -12 X. The four-probe method is not applicable for such small resistance measurements. Therefore, current decay measurements were carried out here to evaluate the joint resistance, and to confirm the formation of the superconducting joint. The details of the measurement method are described in [32]. The measurements were carried out as follows.
First, a single-turn loop consisting of superconducting wires with joints is set in a cryostat, and a current is injected into the loop by magnetic induction. A temporal evolution of the magnetic field around the wire, i.e., current decay, is measured using a Hall sensor. Eventually, the joint resistance can be obtained from the current decay constant (= L/R, where L and R are the self-inductance and the joint resistance of the loop, respectively). The loop sample has four joints, i.e., two NbTi wire/HTT Nb-alloy tape joints, and two Nb 3 Sn/HTT Nb-alloy tape joints, which is needed to connect one NbTi wire to one Nb 3 Sn wire with two HTT Nb-alloy tapes. The selfinductance of the single-turn loop was found to be 0.47 lH in this system.

Results and discussion
Superconducting properties of Nb alloys  magnet systems. Further optimization of B c2 and I c needs to be performed. Presumably, B c2 and I c can be furthermore enhanced by the increase in additive content as well as deformation degree. However, this may worsen drawability owing to the increase in solid-solution hardening.
In contrast, although Nb-62.8%Ti can maintain a B c2 value of approximately 1.5 T after heat treatment, the I c -B curve becomes markedly gentle with rapid I c degradation; the I c and n values at 0.5 T decrease to only 37 A and 32, respectively. Figure 6 shows a SEM image and EDS maps of the NbTi filament after heat treatment. As observed, Nb aggregated in the filaments, and Ti and Cu were interdiffused. Presumably, the Nb containing a small amount of Ti determined the B c2 of the NbTi wire. Furthermore, the V-I curves of the wire were found to be broadened, which indicates that the microstructure no longer has strong pinning properties. The NbTi wire annealed at 685°C was not ductile, either.
The addition of 4 at%Ta-1 at%Hf to Nb enhanced B c2 and I c . These enhancements could be attributed to the addition of Hf, which contributes to maintaining the fine microstructure, as shown in Fig. 2. Previous reports show that the addition of Ta to Nb does not increase but rather depresses the B c2 of Nb [18,20]. The recovery temperature of Nb-4Ta is almost the same as that of Nb, which explains that the addition of Ta has little effect on the superconducting properties of Nb.
Nb-5 at%Zr has poor composite deformability with Cu, and therefore wire could not be cold-drawn and there were no transport properties. According to previous studies, the addition of 5 at%Zr to Nb increases B c2 to approximately 2 T [34,35]. However, as the recovery temperature of the Nb-Zr alloy seems to be below 700°C [35], the steep I c -B curve of Nb-Zr should not be maintained after annealing. In the case of W addition, although a steep I c -B curve was obtained, B c2 was approximately 0.5 T after annealing. Hence, based on the above observations, Hf seems to be the most effective additive element for Nb to realize a HTT Nb alloy with a relatively high B c2 .  a superconducting pathway between them. Furthermore, Hf appeared to diffuse into Nb filaments in the Nb 3 Sn precursor wire from the Nb-Hf core. Figure 7e, f shows the SEM images of the interface microstructure between the NbTi filament and Nb-3 at%Hf core without and with annealing at 450°C for 2 h in a vacuum, respectively. Notably, there are some small gaps visible at the interface without annealing, whereas the bonding appears to be improved after annealing. This additional annealing may be performed as needed.

Microstructure at the interface of the joints
Superconducting properties of the joints Figure 8 shows I c of the joint samples as a function of magnetic field, where I c is defined with a dimensionless criterion of 1 lV. The voltage taps were soldered at a position of approximately 5 mm above the supporting tube surrounding the joint part. The joint configurations were (A) Nb 3 Sn wire/Nb-3Hf tape, (B) NbTi wire/Nb -3Hf tape, and (C) Nb 3 Sn wire/ Nb-3Hf tape/NbTi wire. The Nb-3Hf tapes for joint B were annealed at 685°C before joining. Joint C was a completed form of the joints of NbTi and Nb 3 Sn wires using an HTT Nb alloy intermedium. The magnetic field was applied parallel to the pressed joint surface. The error bar indicates a measured data range. All the joint parts were not stabilized with superior electrical conductors such as Cu. Therefore, the quench phenomenon at more than 100 A, which is a peculiar phenomenon to superconductors, easily damaged the wire. Prevention of the quench phenomenon is expected to be an active area of research in future. Joint A exhibits a steep I c -B curve from 1.15 T with decreasing magnetic field, and I c at 0.8 T surpasses 150 A. Importantly, the yield rate of the successful superconducting joints between Nb 3 Sn wire/Nb-3Hf tape was 100%, and the scattering of the I c data of the joints was found to be extremely small. Even if some filament damage occurred in the Nb 3 Sn precursor wire during the severe cold-pressing, the damage could be repaired as the Nb 3 Sn layer formed at more than 650°C. These findings indicate that superior metallurgical bonding was achieved between the Nb 3 Sn filaments and Nb-3Hf alloy core.
Moreover, some of the joint B specimens also showed a very good I c -B curve. The best I c -B property was superior to that of the annealed Nb-3Hf tape as shown in Fig. 5. Presumably, this I c -B is determined by the property of the Nb-3Hf tape of the joint sample but not the joint part itself. The superconducting properties of the Nb-3Hf tape of the joint were influenced by the magnetic field direction to the tape. In the transport test for the tape, the field was applied perpendicular to the tape axis, while in the joint sample, the field was applied diagonally to the tape axis. In the diagonal field, the Lorentz force working on magnetic fluxes inside are weakened, which results in an increase in I c .
However, a few samples of joint B showed low I c characteristics. Excessive pressure, stress concentration at an edge of the pressing fixture, or filament crossing in the cold-pressing could be the possible reasons, which could cause irreversible severe filament damage. In order to improve the yield rate of the high quality joint, optimization of the coldpressing method and its pressure and the joint length are required.
Some of Joint C samples also showed good I c -B properties, where the best I c value reached more than 150 A at 0.8 T. However, the I c data scattered well similar to joint B. This scattering could be ascribed to the joint property between the NbTi wire and Nb-3Hf tape, considering the small scattering in the superconducting properties of Nb 3 Sn and Nb-3Hf joints.  Figure 9 shows one of the decay curves and the curve fitting using a decay function (V Hall = A ? B exp (-(R j /L) t, where V Hall , R j , L and t are the Hall sensor output, joint resistance and self-inductance of the loop, and A and B are arbitrary constants) at 4.2 K and 0.5 T for a loop sample with NbTi wire/Nb-3Hf tape/Nb 3 Sn wire joints. The external magnetic field was applied parallel to the pressed joint surface. The initial induced current was 47 A. The current decreased rapidly at first and became stable after 1000 s. The curve was fitted between 1000 and 3000 s to evaluate R j . The residual loop current after stabilizing was estimated to be approximately 16 A from the Hall sensor output. The estimated joint resistance was approximately 6.55 9 10 -13 X. The joint resistances at the other magnetic fields were also in the order of 10 -14 -10 -13 X in magnetic fields of 0-0.9 T, which further confirms the superconducting state of the joints.

Joint resistance
The small residual loop current is ascribed to the quality of the joint. After the current decay measurement, the loop sample was cut and the I c -B characteristics of each joint part were measured by a four-probe method. Evidently, the I c of the NbTi/ Nb-3Hf joint showed small values, approximately 16 A at 0.5 T, which corresponds to the residual loop current. This indicates that the induced loop current was limited by the NbTi/Nb-3Hf joint. Moreover, we gave additionally a stronger pressure up to 160-230 kN for the NbTi/Nb-3Hf joint parts, and remeasured their I c . However, I c characteristics were further deteriorated. The pressure of 120 kN may have been too strong for joining of NbTi/Nb-3Hf.

Conclusions
The author developed superconducting Nb alloy intermedia whose critical current does not decrease even after exposure to high temperatures. Based on this Nb alloy, a new concept for metallurgically realizing completely Pb-and Cd-free superconducting joints between NbTi and Nb 3 Sn wires was proposed and proven experimentally. This superconducting joint method is expected to contribute to the realization of environmentally friendly NMR magnet systems.
Among transition elements close to Nb, Hf is the most suitable additive for realizing an HTT superconducting Nb-alloy intermedium with a relatively high critical magnetic field.
A superconducting joint between Nb 3 Sn filaments and one end of a Nb-Hf alloy core was created by forming a superconducting Nb 3 Sn layer at the interface through a chemical reaction. The joint between the NbTi filaments and the other end of the Nb-Hf alloy core was formed via cold pressing to connect their active new surfaces to each other. Ultimately, a superconducting joint between NbTi and Nb 3 Sn wires was realized. The realization of the superconducting joint was confirmed by ultra-low resistance measurements using a current decay measurement system.
The B c2 value of the Nb-3 at%Hf alloy was 1.15 T even after Nb 3 Sn layer formation at 685°C. B c2 needs to be further optimized. The I c value of the NbTi wire/Nb-3Hf tape/Nb 3 Sn wire joint with a joint length of 1 cm at 0.8 T reached more than 150 A. The yield rate of the successful superconducting joints between Nb 3 Sn wire/Nb-3Hf tape was 100%, and the scattering of the I c data of the joints was small. In contrast, I c of the NbTi/HTT Nb-alloy joints shows a relatively larger scattering tendency at the moment. One of the possible reasons for the I c degradation would be an excessive pressure, causing filament damage. Optimization of the cold-pressing method and its pressure and the joint length are therefore necessary to improve the yield rate of high quality joints, especially for NbTi/HTT Nb-alloy joints.