Optimisation of the processing parameters for the fabrication of high-quality joints between Y–Ba–Cu–O single grain, bulk superconductors

High-strength permanent magnets are essential for a wide range of technologies, including levitation devices, motors, generators and magnetic separators. Replacing permanent magnets with single grain, bulk superconductors will enable a step-change in the performance of these technologies by providing an order-of-magnitude increase in magnetic field. However, there remain many key challenges to the practical implementation of bulk superconductors, of which size and geometry are the most fundamental. The current limits to the size and geometry of (RE)-Ba–Cu–O single grain, bulk superconductors would be overcome substantially by the ability to fabricate high-quality joints between these technologically important materials. In this work we present new insights into the creation of superconducting joints between single grain bulk YBCO superconductors using a YBCO-Ag intermediate composition. We have investigated the effect of the joint fabrication temperature on the quality of the joint in order to begin to optimise the joint fabrication route for YBCO. We report on 35 joints produced at different joining temperatures as part of this study. The trapped field properties of the resulting joined samples were measured and the microstructure at each joint was examined. We show that this simple and rapid joining technique is robust to small changes in joint fabrication temperature and suggest routes to further optimise this potentially transformative technique.


Introduction
Permanent magnets are essential to a wide range of modern technologies.The saturation magnetisation, and hence the * Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.performance, of permanent magnets is an intrinsic material property and is limited fundamentally by the number of unpaired electronic spins per lattice site (which effectively form bound currents) in ferromagnetic materials.Bulk superconductors, on the other hand, rely on macroscopic transport currents, rather than bound currents, to generate magnetic field and, as a result, are able to trap magnetic fields that are significantly higher than those generated by conventional permanent magnets.RE-Ba-Cu-O [(RE)BCO, where (RE) is Y, Gd or Sm and consists of a REBa 2 Cu 3 O 7−δ (RE-123) phase matrix containing RE 2 BaCuO 5 (RE-211) precipitates] single grain, bulk superconductors can achieve trapped magnetic fields of greater than 17 T, compared to 1.5 T achieved in the best permanent magnets [1].
Advances in cryogenic technology have now made it feasible to use single grain bulk superconductors as pseudopermanent magnets.This will ultimately enable a step change in the performance of motors, generators, levitation systems, energy storage devices and biomedical imaging equipment, to name but a few applications [2].There remain, however, several key technical challenges to the practical implementation of single grain, bulk superconductors, of which the most fundamental are size and geometry.
It is necessary to fabricate (RE)BCO in the form of large single grains since the presence of high angle grain boundaries in the sample microstructure reduces significantly the electrical connectivity and current flow in multi-grain materials [3].In turn, the trapped field achievable is determined by the product of the magnitude of the current flowing and the area of the current loop it defines [3][4][5][6][7][8].The brittle nature and high hardness of bulk (RE)BCO superconductors makes them difficult to machine and virtually eliminates the possibility of producing highly intricate shapes of these technologically important materials.
Two techniques are used widely for the fabrication of single-grain (RE)BCO: top seeded melt growth (TSMG) and infiltration growth (IG) [6,9].The peritectic reaction between the RE-123 phase and an excess of RE-211 forms the underlying reaction that drives the fabrication process.Both techniques rely on controlled nucleation from a seed and involve extended, carefully controlled periods of slow cooling during the single grain growth process [7].Fabrication is therefore a relatively slow process, taking typically over 4 d to produce a 25 mm diameter YBCO single grain by TSMG and significantly longer for samples of larger diameter.This, coupled with the observation that superconducting properties do not increase proportionally, as would be expected, with increasing sample size, alongside difficulties with cracking and porosity, mean that most bulk single grain superconductors produced are smaller than 30 mm in diameter.Samples larger than 60 mm diameter with good macroscopic superconducting properties are, indeed, rare [10,11].
An alternative approach to fabricating a single, large bulk superconductor is to join a number of smaller, single grains to form one large composite grain with small grain misorientation angles.This would also address the need for complex and conformal geometries required for many applications, including NMR, flywheels and motors and generators [2].
Bulk superconductors fabricated from joined individual grains must fulfil a number of criteria in order to be suitable for practical applications.Most critically, the interfaces should exhibit J c values comparable to that of the parent single grain material.To achieve this, the joints need to approximate to a low angle grain boundary.They should therefore be free of defects, such as trapped liquid phases, pores and voids, and variations in stoichiometry.These factors not only improve the superconducting properties but also enhance the mechanical properties of bulk, single grains.Finally, and equally importantly, the joint must have sufficient mechanical strength to withstand the considerable Lorentz forces experienced by the bulk superconductor both during magnetisation and in service.
The techniques used previously to join single grain YBCO bulk superconductors can be divided into four main categories: solid state diffusion, infiltration joining, welding and silverbased techniques .
Early investigations focussed on solid-state diffusion.In its simplest form, this technique relies on achieving perfect contact between two interfaces and these are joined without the use of additional material.The arrangement is heated without melting and a joint is subsequently formed.This technique requires carefully polished pieces of bulk superconductor to be clamped together at a high temperature for an extended period, while a large load is applied to the grains being joined.Despite the inherent complexities and challenges associated with this technique, a number of studies have successfully produced mechanically robust joints using solid-state diffusion.In all cases, however, the superconducting properties of the joined samples were reported to be notably inferior to those of the as-grown parent single grains [12,13,24,30].
In the second category of technique, infiltration joining, the gap between the two bulk superconductors to be joined is filled with RE-211 powder.An additional powder pellet typically provides liquid phase which, on melting, infiltrates the powder in the gap.This uses the same principle as the IG technique for growing large, individual single grains.The RE-211 powder used has a lower decomposition temperature than YBCO, so the bulk superconductor sections to be joined both remain solid throughout the process and facilitate epitaxial nucleation.Mechanically robust joints with a highly aligned structure and low residual porosity have been produced using this technique.However, to-date no superconducting properties have been reported for this type of joint [31,32].
A third, more widely investigated, set of techniques uses a weld material between the YBCO bulk single grain sections to be joined.The weld material has a lower peritectic temperature than YBCO, so is typically thulium [15,[33][34][35][36], ytterbium [16][17][18][19] or erbium-based [19][20][21] (RE)BCO or YBCO-Ag [22,[25][26][27].On heating, the weld material melts while the majority of the single grain remains solid.The original grain structure is hence retained and acts to self-seed the decomposed weld material.This technique requires an extensive slow cooling period to promote grain growth within the weld composition.The weld material used has taken a variety of forms including powder, a suspension painted onto the surface or pre-sintered green powder compacts.These joints tend to exhibit inferior superconducting properties, which is usually attributed to the accumulation of impurities at the edge of the weld material [33].
The final subset of techniques use a thin layer of a different material, such as silver foil, to reduce the peritectic temperature of the material in the vicinity of the joint interface.The material at the interface subsequently melts while the majority of the YBCO remains solid.One advantage of this method is that the heating profile does not need to be as tightly controlled as in the welding technique, as there is not a thick layer of weld material that must recrystallise during the joining process.Analysis by Iliescu et al suggested that the penetration depth of the silver controlled the thickness of YBCO that melted in this process [37].The resulting joint was of high quality and, encouragingly, produced a trapped field profile that exhibited a single peak [28].This technique yielded a value of J c of 1.35 × 10 4 A cm −2 across the joint, which is comparable to the value of J c observed in a relatively poorquality, single grain YBCO bulk superconductor relative to the current state of the art.
In this paper we extend our previous work that demonstrated the proof of concept of the joining of single grain, bulk YBCO superconductors by verifying extensively, and significantly, that this is a robust and usable technique.We demonstrate further that the use of a single grain YBCO-Ag intermediate material facilitates joining at relatively low temperatures and without the need for a tightly controlled and lengthy slow cooling process.Finally, we have investigated the effect of the temperature on joining, in order to begin to optimise the joint fabrication route for YBCO, in particular.

Sample growth
Thirty five samples of YBCO were grown by liquid-phaseenriched TSMG (LR TSMG) [38].These samples were pressed uniaxially from 99.9% purity powders of Y-123:Y-211:CeO 2 in the mass ratio 150:50:1 in a 20 mm diameter cylindrical die.In addition, 8 samples of YBCO-Ag of 25 mm diameter were grown by LR TSMG [39].The precursor powder was mixed from 99.9% purity powders of Y-123:Y-211:CeO 2 :Ag 2 O in the mass ratio 150:50:1:20.In both cases the liquid-phase-rich powder was mixed from Yb 2 O 3 :Ba 3 Cu 5 O 8 :BaO 2 in the mass ratio 5.0:5.6:1.0.The powder was calcined for 5 h at 850 • C prior to the growth process.These samples were appromimately 10 mm in thickness.
After melt processing, all 35 YBCO samples were annealed in oxygen for a minimum of 8 d at 450 • C.This transformed the Y-123 tetragonal structure to the superconducting orthorhombic phase.The magnetic field trapping ability of the parent YBCO single grain samples could then be measured prior to joining to provide a baseline.

Trapped field measurement
The top and bottom faces of all of the YBCO samples were polished flat and parallel using 180 grit silicon carbide paper.The maximum trapped field at the top and bottom surface was measured initially using a hand-held Hall probe positioned approximately 0.5 mm above the sample surface.Subsequently, the trapped field profile across both the top and bottom surface of each single grain was measured using a rotating array of 19 Hall probes positioned approximately 1.5 mm above the surface of each sample.The samples had been field cooled at 77 K in an applied magnetic field of 1.4 T prior to trapped field measurement.The temperature of each sample was maintained at 77 K for the duration of each measurement.

Joining technique
Each of the YBCO samples was cut in half across a diameter mid-facet line, as shown in figure 1.The two rectangular cross-sections produced by cutting were polished progressively using silicon carbide paper down to 4000 grit.
Slices of thickness between between 1.0 mm and 1.5 mm were cut from the YBCO-Ag single grain samples, as illustrated in figure 1.These slices were polished by hand using 4000 grit silicon carbide paper.
One half of each cut YBCO sample was assembled either side of a single slice of YBCO-Ag, as shown in figure 2. The clamping configuration using a machinable ceramic clamp is also shown in the figure.The samples were loaded under selfweight to aid contact between the two YBCO half samples and the YBCO-Ag slice during the heating process.No further loading was applied to the upper surface of the arrangement.The assemblies were heated in a box furnace in air at a rate of 500 • C h −1 to a range of temperatures, T j , as shown in figure 3 and table 1, held at this temperature for 5 h and then cooled at a rate of 500 • C h −1 to room temperature.
Finally, the joined samples were annealed in oxygen for a minimum of 8 d at 450 • C in order to transform the Y-123 tetragonal structure to the superconducting orthorhombic phase.The trapped field at the top and bottom of the joined samples was measured after the oxygenation process, as described above.

Microstructural analysis
Three joined samples, two joined at 945 • C and one joined at 970 • C, were cut in half along a diameter to expose a central rectangular cross-section.The cut was orientated perpendicular to the joint.The central cross-section was polished using silicon carbide paper of progressively finer grit followed by diamond paste.Scanning electron microscope (SEM) images   were taken at 100x magnification at 20 kV to observe the features at the joint interface.Images were taken at intervals of 2 mm along the c-axis direction of the joined, composite sample.

Sample growth and joining
All 35 of the YBCO samples and all 8 of the YBCO-Ag samples were grown successfully in the form of individual single grains.The measured samples each exhibited a single peak in their trapped field profiles, which is characteristic of a single grain.The maximum fields trapped at the top and base of each of the single grains are shown in figure 4.
A minimum of two samples were joined at each temperature, T j , as listed in table 1. Images of some of the samples are presented in figure 5.Each of the joints fabricated with T j between 950 • C and 985 • C were mechanically robust to manual handling and could easily support their self-weight.In addition, the top surface of each YBCO bulk half-sample retained a single grain appearance.In no case was there obvious signs of large-scale melting of the YBCO at these joining temperatures.The samples joined at 990 • C, 995 • C and 1000 • C did, however, exhibit some signs of large-scale melting, observed by the large number of resulting grains.Joining temperatures above 990 • C are therefore clearly too high to fabricate usefully joined single grain samples of YBCO.
All three samples joined at 930 • C failed to form mechanically robust joints, as did two of four samples joined at 940 • C, and one of three samples joined at 945 • C. This, therefore, suggests that a joining temperature below 950 • C, if held for 5 h or less, is too low to enable joints to be fabricated reliably between YBCO single grain samples.
The maximum trapped field at the top and base of each of the mechanically robust samples is shown in figure 4.An example trapped field profile is shown in figure 6 and the trapped field efficiency of the joints is presented in figure 7. It can be seen from the data in figure 7 that the joining process is reliable at a range of temperatures and so is robust to small changes in joining temperature.
There are number of possible reasons for the degradation in maximum trapped field achieveable between the parent and jointed samples.These include voids, cracks and pores at the joint interface.Another possible, but less likely reason is the diffusion of silver to any gaps at the joint interface to form a non-superconducting intermediate layer.A final possibility is the degradation of the localised J c in the bulk material due to the multiple heating cycles, as discussed in [40].

Microstructure
The microstructure at the joint interface of two samples joined with T j of 945 • C were compared, and the SEM images are shown in figure 8. Sample 945-a achieved 38% of the maximum trapped field of the original sample at the top surface, while sample 945-b achieved 65% of its original value at the same surface.This can be explained by the quality of the joint interface in that the vibrations from the cutting process caused one half of sample 945-a to separate completely from the joint.This suggests that either the joining temperature or time is not sufficiently high to melt sufficient quantities of material in the  vicinity of the joint interface to enable a joint robust to multiple vibration cycles to be produced or that greater loading must be applied to reduce the likelihood of voids remaining present at the interface after joining.In the half of 945-a that did survive the cutting and polishing process, there were a number of voids present at the joint interface when compared to sample 945-b.The joint interface in sample 945-b had an almost continuous joint interface region with the joining interface.It would, therefore, be reasonable to expect that the trapped field properties would be better in sample 945-b than in sample 945-a.
It should also be noted that these samples have experienced two cutting processes and two oxygenation processes.Once the joining process has been optimised there will be no need oxygenate the sample prior to joining as there will be no need to measure the original trapped field.No will there be a need to cut the sample post-joining to observe the microstructure.Therefore, more cracking has occurred than would be observed in a joined sample once the joining technique has been fully optimised.
Variations in the trapped field properties achieved in the joined samples are likely to be due to the additional processing parameters, such as hold time at the joining temperature and loading applied during the joining process.A longer hold time at lower temperatures may enable more material at the joint interface to decompose and hence a more mechanically robust joint with a more continuous interface could be

Conclusions
This study has investigated systematically the effect of the joint fabrication temperature on the superconducting properties of a joined single grain sample, in order to begin optimisation of the joint fabrication process for YBCO.The joined samples have achieved up to 86% of the maximum trapped field achieved in the original unjoined single grain, parent bulk superconductor.This technique is simpler to implement and less time intensive than fabricating larger, single grain bulk superconductors.It is also more robust to small variations in processing parameters.The benefits of a continuous interface at the joint and a single grain intermediate have been observed.The ability to produce composite YBCO single grains reliably using a fabrication route that is robust to relatively large changes in processing parameters provides many opportunities for the widespread integration of single grain bulk superconductors into a range of applications that rely currently on permanent magnets.

Figure 1 .
Figure 1.A schematic illustration of (a) the cut orientation in YBCO bulk material, and (b) the orientation of the slices of joining intermediate cut from single grain YBCO-Ag.

Figure 2 .
Figure 2. The assembly used to join YBCO.

Figure 3 .
Figure 3.The heating profile used to join single grain YBCO bulk superconductors at a range of joining temperatures.

Figure 4 .
Figure 4. Graphs of trapped field post joining (dark grey), prior to cutting and joining (total height of bar).The top graph is of the top of the samples and bottom graph is of the base of the samples, all measured at 77 K after field cooling in an applied field of 1.4 T.

Figure 5 .
Figure 5.A photograph showing an example of one sample joined at each temperature.

Figure 6 .
Figure 6.Examples of the trapped field profile for sample 975-c for: left) before joining, right) after joining.

Figure 7 .
Figure 7. Plots of the joining temperature against the percentage of the maximum trapped field of the original achieved in the joined samples.Top: top surface, bottom: base.

Figure 8 .
Figure 8. SEM images of the joint interface, left) sample 945-a, and right) sample 945-b.

Table 1 .
The joining temperatures used and the number of samples trialled at each temperature.