1. Introduction
High-strength permanent magnets are essential for a wide range of technologies. The saturation magnetisation of a permanent magnet is an intrinsic property of the material, with its performance being limited fundamentally by the number of unpaired spins per site in the crystallographic lattice. Bulk superconductors, which are able to trap magnetic fields significantly in excess of those associated with permanent magnets, however, generate field by the induction of macroscopic currents that flow generally over large sample length scales and are not limited in this way. As a result, rare earth-barium-cuprate [(RE)BCO where (RE) is Y, Gd, Sm or Eu] bulk superconductors can achieve trapped magnetic fields of greater than 17 T [1], compared to 1.5 T achieved typically in the best iron-based permanent magnets [2]. Advances in fabrication processes and cryogenic technology now make it feasible to use bulk superconductors in a range of engineering applications. This will ultimately enable high performance motors, generators, levitation systems, energy storage devices and biomedical imaging equipment to be developed [3].
Applications of bulk superconductors practically require large single grains of (RE)BCO given that the presence of high angle grain boundaries reduces significantly electrical connectivity and current flow within polycrystalline samples. As a result, the maximum trapped field achievable in these materials is determined by the continuous area within a single grain over which a current loop can flow uninterrupted [4], [5], [6], [7], [8], [9].
Two fabrication routes are used widely to produce single grains of (RE)BCO; top seeded melt growth (TSMG) and infiltration growth (IG). This work focuses on samples grown by TSMG, which relies on the peritectic decomposition of a powder compact due to heating above the peritectic temperature. The peritectic temperatures of a range of (RE)-123 phases when heated in air are given in Table 1. The precursor powder decomposes to form a (RE)2BaCuO5 (RE-211) solid phase and a balancing Ba-Cu-O liquid phase composition. Slow, controlled undercooling subsequently enables nucleation and growth to occur from a discrete seed placed usually on the top surface of the precursor pellet prior to heating. If successful, this results a single grain comprising of a continuous superconducting RE-123 phase matrix containing flux-pinning, non-superconducting RE-211 phase inclusions [6], [8].
Table 1. The peritectic temperatures of a range of RE-123 phases measured on heating in air [10]. |
| (RE) component in RE-123 | Peritectic temperature (°C) |
|---|---|
| Nd | 1068 |
| Sm | 1054 |
| Eu | 1046 |
| Gd | 1030 |
| Dy | 1010 |
| Y | 1005 |
| Er | 1005 |
| Yb | 960 |
A single grain bulk superconductors should contain small defects and inclusions distributed uniformly, including the presence of discrete RE-211 particles, in order for its superconducting properties to be optimised. These inclusions and defects should ideally be comparable in size to the coherence length of the Cooper pairs at a given temperature so that they act as effective flux pinning sites [7], [11]. The distribution and size of the RE-211 inclusions, alongside the distribution of silver inclusions, are influenced significantly by the precise TSMG processing conditions. The distribution of RE-211 inclusions within the RE-123 matrix can be modelled approximately by particle pushing and trapping theory [8]. In this, the amount of local undercooling determines the critical radius of the RE-211 particles trapped at each location as the single grain grows. In addition, due to exposure to elevated temperature for extended periods during processing, the RE-211 particles coarsen over time. This can be modelled by the so-called Ostwald ripening theory [8]. The coarsening of the RE-211 particles is governed by both diffusion of solute through the liquid and the reaction at the interface between the Y-211 particle and the liquid-phase [8]. As the peritectic temperatures of Y, Gd, Eu and Sm-123 are significantly different, the fabrication of single grains in each of the systems utilises significantly different heating profiles. The RE-211 and silver inclusions are, therefore, likely to be subjected to a very different set of thermal conditions during the single grain fabrication process.
A higher temperature is necessitated in the initial stages of the heating profile when the RE-123 phase has a higher peritectic temperature to ensure that complete peritectic decomposition occurs. These samples will be exposed to higher temperatures and so there is more heat energy available to allow diffusion within the growing grain than in those where the peritectic temperature of the RE-123 phase is lower. The systems with higher peritectic temperatures do, however, exhibit faster single gain growth rates [12], [13]. As a result, these samples will spend less time at elevated temperatures, and hence there will be a shorter time for diffusion to occur.
As the range of applications for bulk superconductors has grown, interest in the full range of available (RE)BCO materials has increased both for reasons of tailoring materials properties more closly to application requirements and for economic reasons. For example, EuBCO-Ag is prefereable for magnetic resonance imaging applicatons due to reduced paramagnetism and a more homogeneous trapped field, whereas GdBCO has superior mechanical properties and hence is most optimum for very high field applications. YBCO, on the other hand, is more suitable for levitaition applications such as superconducting flywheels [3], [14], [15]. Hence, it may be desireable from an economic perspective as the demand for and cost of specific (RE) elements increases to have multiple options of (RE)BCO available and to understand in detail the implications of varying the (RE)-species in the (RE)BCO composition. This, in turn, will influence critically the fabrication parameters and properties of the resulting material.
In this work we investigate systematically the relationship between the peritectic temperature of different (RE)BCO compositions and the effect this has on the distribution of RE-211 and silver within the single grain microstructure. Four samples were grown by TSMG, one each of YBCO-Ag, GdBCO-Ag, EuBCO-Ag and SmBCO-Ag. These four compositions were chosen specifically since they are the most widely grown types of (RE)BCO single grains. The microstructure and composition close to the top and along the central c-axis of the single grain sample have been studied in detail and their properties and features compared. Conclusions have been drawn on the effect of the peritectic temperature of the system on the distribution and size of RE-211 and silver inclusions for these specific single grains.
2. Method
2.1. Sample growth
One sample each of YBCO-Ag, GdBCO-Ag, SmBCO-Ag and EuBCO-Ag of 20 mm diameter as-grown were fabricated by TSMG [6].
A single grain of YBCO-Ag was grown by liquid-phase enriched TSMG [16]. The precursor powder was mixed from 99.9% purity powders of Y-123:Y-211:CeO2:Ag2O in the mass ratio 150:50:1:20, while the liquid-phase-rich powder was mixed from Yb2O3:Ba3Cu5O8:BaO2 in the mass ratio 5.0:5.6:1.0. Prior to the growth process the liquid-phase-rich powder was calcined for 5 h at 850 °C. The growth process included heating to 943 °C, a hold of 4 h, heating to 1063 °C, holding for 1 h, cooling to 1018 °C at a rate of 50 °C/h, cooling to 1013 °C at 1 °C/h, and cooling to 950 °C at 0.4 °C/h. The sample was then furnace cooled to room temperature.
A single grain of GdBCO-Ag was grown by TSMG. The precursor powder was mixed from 99.9% purity powders of Gd-123:Gd-211:CeO2:BaO2:Ag2O in the mass ratio 150:50:1:2:20. The growth process included heating to 925 °C, a hold of 1 h, heating to 1048 °C, holding for 40 minutes, cooling to 1008 °C at a rate of 75 °C/h, cooling to 998 °C at 0.8 °C/h, cooling to 988 °C at 0.5 °C/h and cooling to 962 °C at 0.2 °C /h. Finally, the sample was furnace cooled to room temperature.
A single grain of EuBCO-Ag was grown by TSMG. The precursor powder was mixed from EuBa2Cu3Ox: Eu2BaCuOy:CeO2:BaO2:Ag2O in the mass ratio 150:50:1:3.5:10. The growth process included heating to 921 °C, a hold of 1 h, heating to 1062 °C, holding for 40 minutes, cooling to 1025 °C at a rate of 75 °C/h, cooling to 1019 °C at 2 °C/h, cooling to 1010 °C at 0.5 °C/h, cooling to 1005 °C at 0.2 °C /h and cooling to 998 °C at 0.1 °C /h. This was followed by furnace cooling to room temperature at a rate of 100 °C/h. Finally, the sample was furnace cooled to room temperature.
A single grain of SmBCO-Ag was grown by TSMG. The precursor powder was mixed from 99.9% purity powders of Sm-123:Sm-211:CeO2: BaO2:Ag2O in the mass ratio 150:50:1:4:20. The growth process included rapid heating to 1058 °C, a hold of 40 minutes, cooling to 1025 °C at a rate of 50 °C/h, cooling to 1015 °C at 1 °C/h, cooling to 1000 °C at 0.3 °C/h, cooling to 995 °C at 0.2 °C/h and cooling to 980 °C at 1 °C/h. Finally, the sample was furnace cooled to room temperature.
All precursor powders were pressed uniaxially using a force of 10 kN in a cylindrical die. The samples were seeded using a thin film seed and melt processed in air. The heating profiles used for each of the samples are shown in Fig. 1.
Fig. 1. The heating profiles used to grow single-grain (RE)BCO samples 20 mm in diameter. The inset amplifies the detail of the region within the circle. |
All four samples were annealed in oxygen after melt processing for a minimum of 8 days at a temperature above 380 °C to transform the RE-123 tetragonal structure to the superconducting orthorhombic phase.
2.2. Microstructural and compositional analysis
Each of the four single grains was cut in half along a diameter midway between the facet lines to expose a central rectangular cross-section. The orientation of the cut is shown schematically in Fig. 2. The central cross-section was polished using silicon carbide paper of progressively finer grit followed, ultimately, with diamond paste to 0.25 μm diamond size. Visible bright field images were taken at 50 times magnification throughout the cross-section to enable a detailed picture of the entire cross section to be constructed from multiple optical images. Images were also taken at 200 times magnification at intervals of 1 mm along the a/b-axis direction at 1 mm from the top of the samples and at 1 mm intervals along the central c-axis.
Fig. 2. The orientation of the cut in each bulk superconductor in plan view. |
In addition, scanning electron microscope (SEM) images were taken at 1000 times magnification at 20 kV to observe the microstructure in more detail. Images were taken at intervals of 1 mm along the a/b-axis direction at 1 mm from the top of the samples and at 1 mm intervals along the central c-axis.
The average composition in each 280 μm by 280 μm approximate area imaged using the SEM was evaluated using an energy dispersive x-ray (EDX) spectroscopy analyser. The composition data were normalized according to the atomic percentages of (RE), Ba, Cu and Ag present within each imaged area (the sum of the wt% of these equals 1) to remove the effect of oxygen. The differences and variation of composition throughout the four samples were compared. Trends in the distribution of RE-211 inclusions were identified using the variation in the calculated ratios of (RE):Ba and (RE):Cu. These ratios correlate directly with the amount of RE-211 present in the samples.
3. Results and discussion
Four single grains were grown successfully as part of this investigation. Photographs of the samples are shown in Fig. 3. Assembled optical microscope images at 50 times magnification are shown in Fig. 4. The bright yellow-white coloured dots incicate silver inclusions, the dark circular regions are pores and the dark lines indicate cracks. The YBCO-Ag sample shows an obvious crack in the mid-section that extends along the a/b-axis direction through most of the sample. The EuBCO-Ag sample shows a similar crack at a similar location. The GdBCO-Ag single grain does not exhibit cracks of this nature, whereas the SmBCO-Ag sample contains a similar crack but which is located towards the bottom of the sample. These cracks are likely to be due to damage by the wall of the die used for pressing rather than being caused by the growth process itself. Silver inclusions are observed throughout the majority of the cross-sectional area in all four single grains investigated and pores are also observed throughout the majority of the cross-section. There are relatively few pores in the outer 5 mm perimeter of each sample as a result of the oxygen produced during peritectic decomposition that is able to diffuse out of the sample at its edge during the growth process. The distribution and size of the pores was studied in more detail using optical microscope images, as shown in Fig. 5.
Fig. 3. Photographs of the single grains of: a) YBCO-Ag, b) GdBCO-Ag, c) EuBCO-Ag and d) SmBCO-Ag. |
Fig. 4. The sample cross sections imaged at 50 times magnification using an optical microscope. a) YBCO-Ag, b) GdBCO-Ag, c) EuBCO-Ag and d) SmBCO-Ag. |
Fig. 5. Microscope images at 200 times magnification taken one third and two thirds of the way along the central c-axis, where the top is just below the seed. The black regions are pores, the white regions are silver, the remainder of the image is the(RE)-123 matrix with small (RE)-211 inclusions. |
The optical microscope images in Fig. 5 were taken at locations approximately one third and two thirds along the central c-axis of each single grain. The locations were chosen in order to provide a good and consistent representation of the sample microstructure but away from the less representative microstructures present towards the edges of the samples. These images show that, in the same sample, there is no noticeable change in the distribution of pores or silver inclusions between the two locations imaged. They do, however, show that there is a difference in the size of the silver inclusions between the different types of sample, with the YBCO-Ag sample containing much smaller silver inclusions and the average silver inclusion and pore size increasing with increasing peritectic temperature.
Detailed composition analysis showed that the normalised weight percentage of silver content along the a/b-axis and 1 mm below the top surface of each single grain varied between 0 and 0.13. No consistent trend was observed in moving from the centre of the single grain towards the edge. The silver content was within a much narrower band of values in the YBCO-Ag sample, followed by the GdBCO-Ag, then the EuBCO-Ag and then, finally, SmBCO-Ag with ranges of 0.01, 0.06, 0.11 and 0.13, respectively. This suggests that a greater variation in silver in each area imaged near the top surface correlates with either a higher processing temperature or with a faster growth rate. This was similar for the c-axis direction, except there was a less obvious trend in the variation with the ranges of 0.09, 0.05, 0.19 and 0.19 observed, respectively. These data are presented in Fig. 6.
Fig. 6. The distribution of silver along top) the a/b-axis at 1 mm from the top and bottom) the central c-axis. |
The greater uniformity in the samples with a slower growth process could be explained by the mechanism of silver migration. The temperature is above the melting point of silver during the TSMG process, which causes the silver particles to melt. The molten silver then moves to fill the pores within the RE-123 matrix. There is a longer period of grain growth for samples with a slower growth process and this, therefore, could lead to a more uniform distribution of pores, leading to a more uniform distribution of silver agglomerates throughout the sample. This can be observed in Fig. 5. The YBCO sample microstructure contains fewer and smaller pores, along with a much more uniform distribution of very small silver inclusions compared with the other single grain microstructures. The distribution of silver and its association with the phase equilibria during the growth of YBCO-Ag has been studied in detail in [17], [18], [19]. These works suggest that when the silver-containing liquid phase has a cpmcentration of silver greater than 12.6 mol% silver inclusions are formed [18].
Composition analysis has also shown that each sample has a very similar distribution of normalised rare earth (RE) element. The amount of (RE) along the a/b-axis increases fractionally within each sample, as predicted by the push-trap theory. Larger RE-211 particles are trapped earlier in the growth process, whereas a larger total volume fraction consisting of smaller average sized individual RE-211 particles will be trapped later in the growth process. The amount of (RE) element is, however, very consistent along the c-axis. The difference in the behaviour along the a/b-axis and the c-axis may be explained by the difference in growth rate along these directions. These data are presented in Fig. 7. The difference in RE-211 distribution along the a/b and c-axis was also was also observed in [20]. It should be noted that the values given for SmBCO-Ag at 10 mm along the a/b-axis and that for EuBCO-Ag at 4 mm along the c-axis are considered to be anomalous results.
Fig. 7. The distribution of (RE) within the samples. Top) distance along the a/b-axis 1 mm from the top, and bottom) along the central c-axis. |
The ratio of RE:Ba and RE:Cu from the composition data were also studied and are shown in Fig. 8. The variation at each location in the different samples was within 0.2 of those for the other (RE)BCO-Ag single grains. This suggests that the distributions of RE-211 inclusions are not affected directly by the maximum temperature used in the fabrication process or by the rate of growth. There is a slight increase towards the edge of the sample along the a/b-axis direction at 1 mm from the top of the sample, suggesting the amount of RE-211 increases with distance along the growth front. This is as expected from push-trap theory and has been observed in EuBCO-Ag in [21] and in GdBCO-Ag in [22]. There is very little variation in the RE/Cu along the c-axis and RE/Ba ratio, suggesting that there is also little variation in the RE-211 distribution in the c-axis direction. As discussed above, the differences in distribution between the two axes may be due to the difference in growth rate, and hence interface energies, as predicted by the push-trap theory, along the respective axial directions.
Fig. 8. Plots of the ratio of (RE)/Cu and (RE)/Ba which correlate with the RE-211 distribution. |
It should be noted that for the push-trap theory [23] to be fully applicable, the inclusions should be fully inert, which is not fully the case in the case of (RE)BCO [24]. This is particularly important in the cases of (RE) = Gd, Eu and Sm, since these (RE) elements can dissolve more readily in the liquid phase and hence the associated RE-211 inclusions are no longer completely inert [8]. This is particularly noticeable for EuBCO-Ag in the data shown in Fig. 7, where the points (with the exception of the anomaly) form a much flatter distribution. Further investigation of the RE-211 particle size and volume fraction could assist with understanding how the RE-211 distribution reflects the push-trap theory in the different materials.
Since the uniformity of the trapped field is influenced by the flux pinning inclusions, both the size and distribution should be as uniform as possible throughout the sample to achieve the best possible performance. These data suggest that the higher temperatures utilised during processing of EuBCO-Ag and SmBCO-Ag do not have a significant impact on the distribution of RE-211 in single grain samples grown by TSMG. This does, however, suggest, in turn, that the uniformity of the distribution of silver is affected by either the higher temperatures utilised during processing or by more rapid growth rates. In order to achieve better mechanical properties, which ultimately influence the absolute maximum trapped field achievable before fracture, the silver inclusions should be distributed uniformly and ideally of a uniform size. As a result, there is scope to further optimise the processing parameters used in the fabrication of EuBCO-Ag and SmBCO-Ag to improve the uniformity of the silver distribution throughout the sample. It is likely that this would also further improve the uniformity of the mechanical properties of the individual single grains, in addition to optimising of the size and distribution of pores.
The combination of the optical microscope images and the composition data suggests that there is no obvious detrimental effect to the RE-211 distribution caused by the high temperatures required to enable peritectic decomposition in the fabrication process of (RE)BCO-Ag single grain supercondiuctors, such as EuBCO-Ag and SmBCO-Ag. The silver distribution is, however, more uniform in the a/b-axis direction in the (RE)BCO-Ag samples requiring lower processing temperatures or having slower growth rates and hence these single grains are generally more likely to exhibit better mechanical properties overall.
4. Conclusion
The maximum overall temperature utilised for the fabrication of (RE)BCO-Ag bulk superconductors was shown to have a limited impact on the distribution of silver and RE-211 single grain samples of YBCO-Ag, GdBCO-Ag, SmBCO-Ag and EuBCO-Ag. This suggests that other factors, such as the cooling rate and compositional refinement additives may have a greater influence on the distribution of these inclusions within the large, single grain microstructure. The superconducting properties of these materials are therefore not limited, or even affected significantly, by the peritectic temperature of the nature of the rare earth element used. This discovery, therefore, provides a greater degree of freedom to use a variety of (RE) materials for a specific economic or functional reason without the need for detailed consideration of the effect of the peritectic temperature of the rare earth material in question.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Additional data related to this publication is available at the University of Cambridge data repository [https://doi.org/10.17863/CAM.99102]. https://doi.org/10.17863/CAM.99102]..All other data accompanying this publication are directly available within the publication.
For the purpose of open access, the author has applied a Creative Commons Attribution (CC BY) licence to any Author Accepted Manuscript version arising from this submission.
Acknowledgements
The authors would like to acknowledge support from the Engineering and Physical Sciences Research Council (EPSRC) grant EP/T014679/1.
