Top-seeded infiltration growth of (Y, Gd)Ba₂Cu₃Oy bulk superconductors with high critical current densities

All the samples have been grown keeping all process parameters the same. Please remember that bulk YG-123 is, in fact, the composite of YG-123 and YG-211. The temperature dependence of dc magnetic susceptibility was measured for all the bulk YG-123 samples (Y80Gd20, Y75Gd25 and Y70Gd30) and compared with the sample Y-123 having only Y-211 inclusions (Y100). Note that the YG-211 particles in the present study are not treated with any grain refining additives like Pt, CeO₂, or PtO₂ [22–25].

Figure 2(a) shows the temperature dependence of normalized dc magnetic susceptibility in the presence of 10 Oe for all the four samples. It is clear from the figure that the superconducting transition temperature T_c³ and transition width ΔT_c⁴ are different for all four samples, although they all show a sharp transition. The samples Y70Gd30 and Y75Gd25 show high onset T_c⁵ to the value of 91.9 K, while the sample Y80Gd20 shows onset T_c of 91.6 K. These values are higher than 91.5 K which is the onset T_c for the sample Y100.

Zoom In Zoom Out Reset image size

Table 2 shows the values of the T_c and onset T_c for all four samples. Sample Y100 shows a T_c of 91.2 K which is rather low compared to the 93.2 K reported by M Muralidhar et al [4] where ball-milled Y-211 of nm order size are used. In our samples, as the content of Gd-211 in the precursor is increased, onset T_c values increased (see the inset in figure 2(a)) accompanied by a reduction of ΔT_c to 0.22 K as seen in sample Y75Gd25 compared to a ΔT_c of 0.52 K in sample Y100 as shown in figure 2(b).

Table 2.  Details of the superconducting transition temperature (T_c) for the various YG-123 (IG) processed samples.

Sample	${{{T}}_{{\rm{c}}}}^{{\rm{onset}}}({\rm{K}})$	Tc (K)
Y100	91.5	91.2
Y80Gd20	91.6	91.5
Y75Gd25	91.9	91.6
Y70Gd30	91.9	91.5
Y75Gd25 B1	91.9	91.6
Y75Gd25 B2	91.6	91.0
Y75Gd25 B3	91.0	90.8
Y75Gd25 S1	91.9	91.6
Y75Gd25 S2	91.6	91.3

Figure 2(c) shows J_c—H curves for YG-123 samples with different YG-211 contents at 77 K with field applied parallel to the c-axis calculated from the M-H curves using the extended Bean model formula [19, 20]. The J_c values of sample Y100 at 77 K and 0 T for the H∣∣c-axis was 3.71 × 10⁴ A cm⁻², which is relatively high without any treatment of Y-211 particles [26] and is the benefit of the IG process. The J_c at 77 K and 0 T for sample Y75Gd25 showed the highest value of 4.78 × 10⁴ A cm⁻² as compared to the values of 3.80 × 10⁴ A cm⁻² and 4.18 × 10⁴ A cm⁻² for samples Y70Gd30 and Y80Gd20 respectively. However, in intermediate fields for 1.5 to 3 T, sample Y70Gd30 exhibited J_c values higher than those of Y75Gd25 and Y80Gd20. The critical current densities survived until μ₀H = 4.03, 4.49, 4.84 and 5.10 T for samples Y100, Y80Gd20, Y70Gd30 and Y75Gd25, respectively. These values correspond to the irreversibility field, B_irr (i.e. the field at which J_c = 100 A cm⁻²), that increased with increasing Gd-211 content due to the higher pinning efficiency of Gd-211 inclusions compared to Y-211. The maximum B_irr for Gd-123/Gd-211 composite is reported to be ∼5.5 T at 77 K [27]. Clearly, as the Gd-211 content in the precursor is increased, the J_c values increase. It is notable that top-seeded IG-processed samples exhibit high critical current densities without any refinement of RE-211 particles, unlike TSMG process where the refinement of RE-211 particles with Pt and/or CeO₂ additions are required for higher J_c values. Delamare et al [28] and Shi et al [29] have reported J_c values of 5.5 × 10⁴ A cm⁻² and 7.2 × 10⁴ A cm⁻², respectively, for melt-textured Y-123 and Gd-123 samples with the secondary phase refinement additives.

In order to check the spatial variation of critical current density J_c and superconducting transition temperature T_c throughout the Y75Gd25 sample, we measured the T_c values for the samples cut from various locations as shown in figure 3. Figure 4(a) shows the temperature dependence of normalized dc magnetic susceptibility in the presence of a magnetic field of 10 Oe. Samples B1 and S1 (i.e. the parts cut out from closer to the seed) show a sharp transition with high onset T_c of 91.9 K, while samples B2 and S2 exhibit onset T_c of 91.6 K and sample B3 shows the lowest onset T_c value of 91 K. The values of the transition temperatures are presented in table 2, which shows that there is a spatial variation in T_c throughout the sample ranging from 91 K to 91.9 K, which is a feature of IG-grown samples as also reported by Iida et al [30].

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Figure 4(b) shows the field dependence of J_c values at 77 K for H∣∣c-axis calculated from M-H loops based on the extended Bean model for various positions in the Y75Gd25 sample. The plot shows that B1, B2, B3, S1 and S2 has J_c values of 4.78 × 10⁴, 4.70 × 10⁴, 4.10 × 10⁴, 3.80 × 10⁴ and 3.70 × 10⁴ A cm⁻², respectively at 77 K in zero field, indicating that there is also a spatial variation in J_c values in the sample. The J_c values also decreased when increasing the distance from the seed along both directions of a- and c-axes, which is similar to the report by Iida et al [31] in both IG and melt-textured YBCO samples.

In order to clarify the pinning mechanism of the YG-123 samples, we analyzed the global volume pinning forces F_p utilizing the scaling theory [32, 33] where F_p = J_c × B_a from the critical current densities J_c (H,T) and applied field B_a. The peak in the plots of F_p versus μ₀ H is F_p,max. We have then plotted F_p/F_p,max versus reduced field h = H_a/H_irr , where H_irr is the irreversibility field. Figures 5(a) and (b) show the plots for the sample Y100 and YG-123 bulk samples. The data were then fitted with a general expression of pinning function [32, 33] :

$$\begin{eqnarray}&&\displaystyle \frac{{F}_{{\rm{p}}}}{{F}_{{\rm{p}},{\rm{\max }}}}=A{h}^{p}{(1-h)}^{q}\end{eqnarray} \tag{ 2 }$$

where the parameters p and q denote the characteristic of pinning in the sample and the peak of the function is given by h_max = p/(p + q). Clearly in figure 5(a), the plot for IG-processed Y-123 with Y-211 in the precursor shows a peak at h_max = 0.33 which gives p = 2 and q = 4 with A = 45.8. Similar values have been reported by J N Li et al for single crystalline YBa₂Cu₃O₇ [34] where both the pinning and shear strength are of comparable level, leading to a higher exponent q value [35]. This suggests that normal and point pinning is active as in melt-processed Y-Ba-Cu-O containing fine Y-211 inclusions [36]. Similarly, h_max = 0.33 has also been reported in TSMG-processed NdBa₂ Cu ₃O_7−δ with p = 1 and q = 2, suggesting that normal and point pinning is active based on small size normal cores [37].

Zoom In Zoom Out Reset image size

Figure 5(b) shows the plots for the samples Y80Gd20, Y75Gd25 and Y70Gd30. The peak position of the samples YG-123 shows h_max = 0.40. Clearly the peak at h_max = 0.40 is similar to reports of 0.48 and 0.42 for single crystals and OCMG-processed Nd-Ba-Cu-O samples [38], and also similar to reports by A K Pradhan et al [39]. Compared to IG-processed Y-123, the Gd-containing YG-123 bulk samples exhibit more effective pinning and hence higher h_max. This also implies that the presence of δT_c pinning is active in the sample due to variation in the distribution of YG-211 distribution (compositional fluctuations leading to fluctuations in transition temperature T_c), or this can also occur due to variation in oxygen distribution within the sample.

Figure 6(a) shows the FESEM micrograph image of sample Y100 at 2500x magnification. The Y-211 inclusions of micron order are clearly visible and uniformly distributed throughout the Y-123 superconducting matrix. To confirm its chemical composition, energy dispersive x-ray spectroscopy (EDS) point analysis was performed for two different 211 inclusions and found to have the element ratio of Y:Ba:Cu = 2:1:1 as shown in figure 6(b). The FESEM micrographs observed at 2500x magnification for the bulk YG-123 samples are presented in figure 7 along with YG-211 particle size distribution histograms in panels (d)–(f). Micron-sized YG-211 particles (light gray) are distributed almost uniformly in the YG-123 matrix (dark gray). FESEM observations for the bulk samples Y80Gd20, Y75Gd25 and Y70Gd30 suggested that the average size of the YG-211 particles was reduced by increasing Gd-211 content. Table 3(a) contains the volume fraction of YG-211 particles $({{V}_{f}}^{211})$ and the average size of YG-211 particles: d (in μm). Thus, the reduction in YG-211 particle size for the bulk Y80Gd20, Y75Gd25 and Y70Gd30 samples is also evident from table 3(a). Along with the size reduction, the volume fraction of these YG-211 particles also decreased significantly. Similar results have been reported by Hinai et al [40] and also by Iida et al [31] where the volume fraction of Gd-211 particles trapped by Gd-123 matrix decreased when increasing Ba content in the liquid source pellet.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 3.  Details of the percentage volume fraction and YG-211 particle size from the FESEM microstructure analysis.

	Sample	${{V}_{f}}^{211}$ (%)	Size d (μm)	${{{\rm{V}}}_{{\rm{f}}}}^{211}/{\rm{d}}$ (μm−1)
(a)	Y80Gd20	39.14	1.90	20.60
	Y75Gd25	34.59	1.65	20.96
	Y70Gd30	32.69	1.57	20.90
(b)	Y75Gd25 B1	34.59	1.65	20.96
	Y75Gd25 B2	30.44	1.55	19.64
	Y75Gd25 B3	20.19	1.28	15.77
	Y75Gd25 S1	31.08	1.41	22.04
	Y75Gd25 S2	29.49	1.33	22.17

The IG process does not involve the peritectic decomposition of RE-123 to RE-211. Instead, it involves the infiltration of the liquid phase into the solid RE-211 precursor and subsequent solidification in the presence of a continuous supply of the liquid, leading to the formation of the RE-123 phase. The growth of the RE-123 phase continues with the supply of RE ions from the RE-211 particles in the RE-123 matrix through peritectic temperature T_p and thus RE-211 dissolves while excess RE-211 is trapped in the matrix [41]. The peritectic solidification of RE-123 while slow cooling from RE-211 can be understood via three steps: (a) growth of faceted 123 crystals, (b) melting of 211 particles in liquid and (c) solute diffusion in liquid [42]. The RE-211 in the presence of excess liquid (BaCuO₂ + CuO) transforms into RE-123 along the liquidus lines of RE (Y, Gd) in Ba₃Cu₅O₈ melt in air [43]. On slow cooling from T_p, the growing RE-123 will trap the RE-211 in liquid matrix according to the reaction: RE-211 + L → RE-123 + L. Thus, the RE-123 phase precipitates, the liquid becomes RE deficient and the required RE flux is provided by the simultaneous dissolution of RE-211 particles. The growth front thus not only dissolves the RE-211 particles but also traps RE-211 particles within 123 matrix depending upon the dissolution rate of RE-211 particles and rate of growth of the RE-123 matrix. This can also be understood in terms of the fact that the RE solute diffusion occurs from composition difference at the liquid RE-211 and liquid RE-123 interface and is also related to the growth rate of RE-123 phase [44]. Since RE-211 particles dissolve in liquid to form RE-123, each RE-211 particle thus becomes smaller according to the amount of RE-211 being used up [45]. Krauns et al [43] give the liquidus lines of Y and Gd in Ba₃Cu₅O₈ where transformation of RE-211 + L into RE-123 + L occurs at temperature of 1005 °C for Y and at 1045 °C for Gd. Thus Gd has higher T_p, and they have also shown that the enthalpy of dissolution for RE-211 compounds is higher for Gd-211 than Y-211. Therefore, during the heat treatment schedule when all the liquid has melted during the first stage at 1040 °C, the Gd-211 particles with higher T_p start to be consumed before Y-211 with lower T_p and form Gd-123 with the required Gd flux being provided by the Gd-211 inclusions present in the precursor. The growth rate is thus defined by the diffusion rate of RE ions and the more soluble the element is, the faster is the growth process [46]. For sample Y80Gd20 where the content of Gd-211 is less than Y70Gd30, the growth rate is initially smaller until the temperature reaches 1005 °C. Below this temperature, Y-211 starts to dissolve. So, a variation in the distribution of YG-211 particle size can be observed, and it decreases with increasing Gd-211 content. Hence sample Y70Gd30 has smaller sized YG-211 particles due to a higher growth rate above 1005 °C. As a result, the size of YG-211 particles of Y80Gd20 is larger than Y70Gd30, as can also be seen from table 3(a).

Figure 8 shows the plot of critical current density J_c at μ₀ H = 0.2 T and 77 K for samples Y80Gd20, Y75Gd25 and Y70Gd30 against the ratio of volume fraction and an average size of YG-211 particles. This plot shows that J_c linearly varies as ${{V}_{f}}^{211}/d$ as has been reported by Murakami et al [3].

Zoom In Zoom Out Reset image size

Figure 9 shows the FESEM microstructure of various positions (as shown in figure 3) in sample Y75Gd25 at a magnification of 2500x. Table 3(b) shows the volume fraction and average size of YG-211 particles for the various positions in the sample. From the tabulated values and histograms in figures 9(f)–(j), one can clearly see that the average size of 211 inclusions decreased as the distance from the seed increased along both a- and c-axis directions. However, Ouerghi et al [47] have reported an increase of RE-211 particle size as we proceed away from seed due to Ostwald ripening [48], whereby larger sized 211 particles are found at the edges of the sample [49–51]. A decrease in the particle size can well be described by pushing/trapping theory [44, 49, 52, 53], which was however proposed for non-reactive inclusions, but since the 211 particles are consumed during the growth of the 123 phase, the smaller particles pushed out are accumulated near the edges. In addition, these particles are bound to become more smaller due to self-decomposition as the growth front proceeds in both the a-axis and c-axis from the seed crystal [52]. From table 3(b), we can see that the volume fraction of YG-211 particles decreases along both the a- and c-axis directions by increasing the distance from the seed crystal, which is not consistent with the pushing/trapping theory where more RE-211 particles will be pushed away from the seed, and the volume fraction should increase. K Iida et al [31] has reported an increase in ${{V}_{f}}^{211}$ with distance from the seed in IG-processed Gd-123 system. Diko et al [54] have also reported a similar increase in ${{V}_{f}}^{211}$ in melt-grown REBa₂Cu₃O_y. Similar reports were also reported by Endo et al [55] and for nanocomposite inclusions in REBCO by N Hari Babu et al [56]. However, contradictory to all these reports, our samples show a decrease in volume fraction which can be due to the fact that the YG-211 particles being pushed out should also dissolve at the growing 123 front to provide the necessary RE ion flux, thus leading to a significant decrease in volume fraction as proposed by Chow et al [57]. When we treat the trapping/pushing phenomenon, many factors should be taken into account since the RE-211 inclusions are reactive in nature. The slow dissolution of RE-211 particles leads to a reduced volume fraction, which is evident from the histograms of YG-211 particle size distribution in figures 11(f)–(j).

Zoom In Zoom Out Reset image size

Figure 10(a) shows the plot of J_c at T = 77 K and H = 0.2 T versus volume fraction of YG-211 particles for various positions along the c-axis in the sample Y75Gd25. It is evident from the plot that J_c increased with the volume fraction of YG-211 particles in the presence of a magnetic field, which has also been reported in REBCO system by N Hari Babu et al [56]. Figure 10(b) shows the plot of J_c versus the ratio of volume fraction and the average size of YG-211 inclusions for the same sample Y75Gd25. This is also in good agreement with the linear relationship between critical current density and the ratio of volume fraction to the average size of 211 particles, as already established by Murakami et al [3].

Zoom In Zoom Out Reset image size

The EDS analysis was also carried out for the sample Y75Gd25. A very interesting feature has been observed in these YG-211 inclusions; namely, that these inclusions show a core-like feature which by elemental mapping (under a magnification of 15 000x) has been confirmed to be yttrium rich core. Figures 11(a) and (b) show the EDS point elemental analysis of three different points within the same YG-211 particle showing that it contains different concentrations of Y:Gd. This core is almost present in every YG-211 inclusion in the matrix.

Zoom In Zoom Out Reset image size

Figure 12(a) shows the Raman scattering spectrum of an IG-processed Y-123 sample, i.e. for sample Y100, where 140 cm⁻¹ is assigned to CuO₂-plane copper in-phase mode (Cu2), 334 cm⁻¹ (B_1g) mode is CuO₂-plane oxygen out of phase mode (O2+/O3−) and A_g mode near 452 cm⁻¹ is the CuO₂-plane oxygen in-phase mode (O2+/O3+) [58, 59]. These modes do not depend on oxygen content or cation stoichiometry. The 500 cm⁻¹ is the apical oxygen mode (O4) associated with the stretching of the CuO bond along the c-axis and is highly dependent on the oxygen content of the sample [60]. The position of 500 cm⁻¹ mode shifts to 475 cm⁻¹ in tetragonal phase [61]. It varies from 492 to 502 cm⁻¹ for an oxygen content of 6.2 to 7.0 [62]. Information regarding grain alignment is inferred from the intensity ratio of apical oxygen (O4) mode and O2+/O3− mode [61]. The mode at 118 cm⁻¹ is due to Ba atoms. In addition to these modes, two more modes at 248 cm⁻¹ and 596 cm⁻¹ are observed [58], which are assigned to stretching vibrations of copper and oxygen atoms at the end of short chain fragments [63]. Other secondary phase peaks may also appear due to the presence of Y₂BaCuO₅ [64]. The peak at 300 cm⁻¹ is due to CuO [65], and the peaks at 592 cm⁻¹, 563 cm⁻¹ and 633 cm⁻¹ are due to Ba₂Cu₃O_5.9, Ba₂CuO₃ and BaCuO₂, respectively [66].

Zoom In Zoom Out Reset image size

Frequencies of intrinsic Raman modes of REBa₂Cu₃O₇ (RE = Y, Gd) are slightly dependent on ionic radii of rare earth elements (Y, Gd) [67]. However, 452 cm⁻¹ mode of O2+/O3 + and Cu2 mode at 150 cm⁻¹ show no variation with ionic radii in YBa₂Cu₃O₇ and GdBa₂Cu₃O₇. Raman spectra of (Y, Gd)Ba₂Cu₃O₇ are the same as those of other RE-123 superconducting samples. Our IG-processed bulk YG-123 samples show a similar spectrum to Y-123.

Figure 12(b) shows the Raman scattering spectrum of YG-123 bulk samples produced by the IG method. A similar spectrum has also been reported for TSMG-processed bulk Y-123 samples by Delorme et al [68]. The position of O4 mode between 491 and 500 cm⁻¹ signifies that these YG-123 are in their orthorhombic (superconducting) phase.

Figure 12(c) shows the Raman spectrum of the Y-211 and YG-211 inclusions in Y-123 and YG-123 bulk IG-processed samples respectively. The peaks with ω > 300 cm⁻¹ belong to oxygen vibrations [64]. Both the spectra are almost similar and the peak positions match with the modes of Y-211 and Gd-211 as given by Abrashev et al [64]. Jin et al [59] have also reported the similar spectrum of (Y,Gd)BCO thin film samples as our IG-processed YG-123 spectrum. Since both the O2+/O3− and O4 modes are present in the spectrum of YG-123, we can conclude that our IG-processed YG-123 samples have both a-axis as well as c-axis oriented grains.