Influence of Milling Energy and Precursors on CaKFe₄As₄ Polycrystalline Superconductor Morphology

Abstract

The synthesis of CaKFe₄As₄ superconducting compounds either requires the adoption of high-temperature synthesis or implies the intimate mixing of the precursors via mechanochemical routes before the thermal step in order to avoid chemical inhomogeneities that lead to thermodynamically stable unwanted phases. High Energy Ball Milling (HEBM) represents a useful tool to ensure the comminution of the elements and their dispersion to obtain the target phase. The adoption of mechanochemical treatments is, however, known to lead to the formation of aggregates of small crystals, leading to a powder morphology not optimal for practical applications. In this work, we report our findings in the synthesis of CaKFe₄As₄ polycrystalline compounds showing the effect of milling energy on the morphology and phase composition of the powders. To overcome the limits of conventional synthesis, we report the results of a novel synthesis approach for CaKFe₄As₄ materials, highlighting how the choice of the proper precursors and the adoption of milder treatments can represent the key to optimizing the powder morphology.

1. Introduction

Iron-based superconductors (IBSCs) represent a wide class of materials, characterized by the insurgence of superconductivity with critical temperatures up to 60 K, discovered in the late 2000s [1,2]. The interest in this class of materials derives from the large magnetic critical fields, allowing to foresee their application in low-temperature and high-field regimes, and the possibility of using some of the IBSCs to produce superconducting wires with the Powder in Tube (PIT) method. Among IBSCs, CaKFe₄As₄ is a material of the 1144 family, discovered in 2016 by Iyo et al. [3]. It has a structure similar to 122-type IBSC compounds, but with a higher symmetry that leads to a tetragonal unit cell with a space group P4/mmm. In fact, compared to the (Ba,K)Fe₂As₂ compound, where Ba and K occupy the same atomic position, intercalating FeAs planes [4], Ca and K in CaKFe₄As₄, due to their difference in ionic radii, occupy different and alternating planes between the FeAs layers [3]. This material is characterized by a T_c of approximately 35 K, H_c2 higher than 630kOe and low anisotropy γ = 2–3 [5,6]. Among the 1144 compounds, it is considered the best candidate for use as bulk due to the easily available and inexpensive chemical constituents and its high critical currents J_c, both of which are fundamental for large-scale applications. In particular, the J_c of Ca/K-1144 single crystals, in the order of 10⁶ A/cm² in self-field, is high and robust against temperature and field [5,6,7]. For these reasons, CaKFe₄As₄ has been evaluated as a promising material for the production of IBSC wires [8,9].

In polycrystalline materials, however, the value of the critical current is highly affected by the quality of the microstructure of polycrystalline samples. In particular, in IBSCs, the critical factors turn out to be impurities at the grain edge, porosities and large misalignment of superconducting grains [10]. Therefore, it is of primary importance to increase the quality of the microstructure of the polycrystalline material in order to improve the value of the intergrain transport current. Grain size can play a crucial role in this, as larger particles allow us to obtain an efficient texture during the mechanical deformation of the wire [11,12,13,14]. Synthesis of CaKFe₄As₄ can be carried out at high temperatures (e.g., close to or exceeding 900 °C) [3,15,16] or at lower temperatures, by introducing a preliminary High Energy Ball Milling treatment (HEBM) [17], as other IBSC compounds [18]. A HEBM treatment influences the powder morphology, promoting the formation of nanocrystalline and, potentially, aggregates [19]. The fine control of the powder morphology depends thus on the energy transferred to the powders during the process. At the same time, a sustained mechanochemical process, which is not simple blending, is necessary to achieve a sufficient degree of mixing of the elements to avoid inhomogeneities during the high-temperature treatment [20].

In this work, we show how the milling process influences the morphology of CaKFe₄As₄ polycrystalline compounds. Furthermore, in order to be able to adopt milder HEBM treatments to enhance the grain size with respect to our previous works, we developed a novel synthesis approach for CaKFe₄As₄. This consists of the substitution of the bulk metallic calcium with calcium hydride powders in order to be able to disperse it more efficiently, similarly to as previously reported for (Ba,K)Fe₂As₂ [21].

2. Materials and Methods

The material synthesis was carried out similarly to as previously reported [22]. All the material handling was carried out in an Ar glove box. Briefly, the precursors (arsenic powder from Alfa Aesar, 99%, potassium cubes from Alfa Aesar, 99.5%, calcium granules from Sigma Aldrich, Milano, Italy, 99.5%, iron powder from Alfa Aesar, 99.998%, calcium hydride from Sigma Aldrich, 99.9%) were weighed as indicated in

Table 1. Summary of sample composition and milling parameters.

Sample	Milling Energy	Nominal Composition Ca:K:Fe:As	Ca Source	Ball Diameter (mm)	Ball to Powder Ratio	Milling Time (h)
S1	2.3 MJ/g	1.27:1.18:3.8:4	Ca metal	12	30	5
S2	0.9 MJ/g					2
S3	0.2 MJ/g	1.35:1.05:3.5:4	CaH2 powder	8	15	1
S4	0.1 MJ/g					0.5

, mixed in a 60 mL stainless steel grinding jar and milled in a SPEX8000M mixer mill. The nominal chemical composition is the result of experiments carried out in order to balance the commercial precursor chemical composition, as described previously [20]. After the milling process, a homogeneous, fine black powder is obtained. This is pressed into disk-shaped pellets (diameter d ≈ 5 mm, thickness h ≈ 0.5–1 mm) by uniaxial pressing and heated inside a sealed steel tube containing a graphite crucible for 10 h at 800 °C with a 10 °C/min heating rate and a 5 °C/min cooling rate. Pellets with an average statistical geometric density of approximately 80% were obtained after this step.

The energy transferred to the powders was calculated as the integral over the time of the energy transferred by a single hit, considering the frequency and the number of balls, assuming a ball velocity for this experimental set-up of 4.2 m/s [23].

The morphology of the powders was studied on the fractured pellets to observe the bulk morphology using SEM/EDS analyses with a Leo 1515 SEM equipped with an Oxford x-act EDS system. XRD measurements were carried out on sealed capillaries using a Smartlab Rigaku with CuKα radiation in the 10–110° 2θ range.

Magnetic measurements were performed using an Oxford Instrument Vibrating Sample Magnetometer (VSM). All magnetic measurements were carried out on the above-mentioned sintered disks with the field applied perpendicularly to the disk surface, and the results were divided by the geometrical volume to obtain the Magnetization M. For the calculation of the shielded volume, geometrical demagnetization factors were calculated in the disk shape assumption as N_z = 1 − πh/d, where h and d are, respectively, the thickness and diameter of the disk [24]. Then, the experimental magnetic susceptibility χ_exp = M/H was calculated to assess the magnetic susceptibility χ = χ_exp/(1 − z·χ_exp). The critical current density was estimated by applying Bean’s model [25] from the magnetic hysteresis loops as J_c = 30·ΔM/d, where ΔM is the difference between the upper and lower branches.

3. Results

In Figure 1, we report the micrographs of the different samples.

Starting from panels (a) and (b), referring to sample S1—subjected to a 2.3 MJ/g milling treatment—from the low magnification image, it is evident that the compound is formed by large micron-sized aggregates. The higher magnification image allows us to observe the finer structure, where the aggregates are formed by large amounts of tens of nanometer-sized crystals, with fewer larger crystals grown in the open porosities as shown in previous syntheses [17,20].

Sample S2, produced with a milder ball-milling energy (E_BM = 0.9 MJ/g) treatment of the pure elements, is depicted in panels (c) and (d). The material is formed by coarse particles, but the appearance is smoother with respect to the more energetic treatment, suggesting a lower degree of aggregation. Observing the high-magnification image, it is evident that the grains are more structured and clearly resemble flat platelets, as expected from the layered IBSC structure. The reduction in the energy involved therefore seems to have a beneficial effect on the morphology of the produced samples.

Adopting lower milling energies when using the metallic calcium precursor leads to retrieving an inhomogeneous powder composed of fine black particles and coarse metallic chunks of calcium. Therefore, the adoption of an energetic process, as previously introduced, is needed in order to comminute calcium and obtain a homogeneous powder.

The adoption of the CaH₂ powder precursors allowed us to instead obtain a homogeneous powder for milder treatments. In panels (e) and (f), the images of sample S3, processed with 0.2 MJ/g milling energy, are reported. The morphology of the material is clearly different from the previous examples, as evident already from the low-magnification images. The compound is formed by fine-structured grains that are hundreds of nanometers in size, and several larger crystals that surround the smaller particles. In our previous works, we related the appearance of large crystals with the presence of open porosity, which may allow for vapor phase transport mechanisms and promote this kind of growth. The abundance of these large particles may thus suggest that a highly porous structure is formed during the 800 °C temperature treatment.

The high magnification image shows the finer structure, with several well-structured and faceted platelet-like grains randomly oriented. The evolution of morphology with respect to the 0.9 MJ/g sample is remarkable, with larger grains in a less compact arrangement.

Finally, the morphology sample S4 is reported in panels (g) and (h). In line with the observed trend, the material seems formed mostly by micron-sized grains, even if smaller particles are evident in the high magnification image, suggesting that the nucleation and growth phenomena that drive the morphology of the material may still be influenced by the mechanochemical process despite such a low energy milling treatment.

To semi-quantitatively evaluate the sample morphology, the average grain size is reported in Figure 2 as a function of the milling energy. As expected, we can observe a decreasing trend, with values close to 1 µm for low-energy milling treatments that approach 50 nm for the high-energy process. The results are coherent in terms of grain size and corresponding energy values with other IBSC compounds, as reported by Tokuta et al. in a Ba(Fe,Co)₂As₂ polycrystalline bulk sample [26]. In this case, it is shown how the morphology of the powders varies between hundreds of nanometers and tens of nanometers when the energy is changed from 20 kJ/g to 0.5 MJ/g. The slight difference with our case can be ascribed among the other factors to the different thermal processes. Furthermore, Ba-122 compounds is characterized by a mechanochemically-activated self-sustained reaction that may influence the nucleation mechanisms during the high-temperature step and, thus, its grain size [18,26].

To assess the phase composition of the compounds synthesized by the hydride route, XRD measurements were carried out. The results are reported in Figure 3. The main phase for both compounds is the Ca/K-1144 phase, as evident from the comparison with the simulated spectra [3]. The pattern of the sample produced by the milder HEBM treatment, however, also shows the presence of several impurities. Peaks ascribable to Fe, Fe₂As, and K-122 phase are highlighted in the figure, suggesting that the degree of mixing of the elements is not sufficient to ensure the production of a homogeneous phase during the following thermal treatment. The sample processed with 0.2 MJ/g, instead, shows structural features similar to the samples produced by the elements route, which is the vast majority of the 1144 phase, and minor peaks due to the presence of calcium oxide [20,22].

For this reason, further analysis to qualify the novel synthesis was carried out focusing on sample S3. The chemical composition was evaluated by means of SEM/EDS measurements on a compacted sample, and the averaged result is reported in

Table 2. Chemical composition (metal atomic %) of the S3 (hydride route) sample as measured from SEM/EDS.

Element	At %
Ca	10 ± 1
K	11 ± 1
Fe	41 ± 2
As	38 ± 4

, showing good agreement with the XRD result and the expected chemical composition, which is Ca:K:Fe:As = 10%:10%:40%:40%. It is worth noting that the measured chemical composition is quite different from the nominal one (CaH₂:K:Fe:As = 13.5%:10.5%:35%:40%). This is due to the initial contamination of the commercial reactants, which causes a large deviation between the final product and the nominal composition [22] and potentially leads to the loss of volatile elements, which should, however, be minimized at 800 °C [22].

The resistance as a function of temperature for sample S3 is reported in Figure 4. The curve is characterized by a monotonous decrease in resistance when lowering the temperature, with an initial smoother reduction that reaches a linear behavior for temperatures below 150 K. A Residual Resistivity Ratio factor of approximately 7, calculated as R_T=273K/R_T=40K, is, to the best of our knowledge, among the highest values found in the literature for similar polycrystalline compounds [15,16,22] and can be associated with good quality of the sample. In the inset, the superconducting transition is highlighted: below 36 K, the resistance drops abruptly, reaching zero at approximately 34 K. Also, this value constitutes a good benchmark with respect to previous results of polycrystalline samples not processed at high pressures, where onset values in the 33–35 K range are commonly observed, and zero resistance values generally vary between 30 K and 33 K [15,16,27,28].

The superconducting transition is also shown in Figure 5a by means of magnetic measurements. As evident from the figure, at low temperatures, the sample exhibits a strong diamagnetic screening. Values of magnetic susceptibility close to −1 are, in fact, obtained from the measurement for temperatures close to 5 K. The diamagnetism tends to vanish at temperatures close to the critical temperature, i.e., 35 K, where the curves tend to zero. The temperature at which the ZFC and FC curves split can, instead, be associated with the irreversibility temperature, with values in good agreement with what is observed by resistivity measurements and in line with previous reports [15,16,27,28]. Despite a sharp resistive transition and a good phase homogeneity, the magnetic transition shows, however, a certain degree of broadening. This can be ascribed to the relatively small crystallite size, which can generate a broad dependence on the temperature of the diamagnetic screening due to the effective shielding changes related to the variations of the London penetration depth, as also shown recently for similar polycrystalline IBSC compounds [29].

The magnetic hysteresis loops measured at 4.2 K, 10 K and 40 K are reported in Figure 5b. It can be observed that the curves obtained at low temperatures exhibit an open hysteresis, which is evidence of a supercurrent flowing in the sample. The shape, however, denotes evidence of asymmetries that can be related to granularity phenomena [30,31], i.e., the presence of obstacles to the flow of the currents, such as the high-angle grain boundaries due to the random orientation of crystallites that are evident from the microscopy analysis. The curve measured above Tc does not highlight a significant background, suggesting that no substantial presence of magnetic impurities resides in the sample. By applying Bean’s model, critical currents in the range 10²–10³ A/cm² were calculated as a function of the field and are reported in the figure inset. The values are lower than those of the densified polycrystalline samples (e.g., [15,16,31]), most likely reflecting a high degree of misorientation and porosity in the synthesized compound. Nevertheless, the produced compound shows promising aspects in terms of homogeneity and improved morphology for the production of superconducting wires via the powder-in-tube method.

4. Conclusions

The results reported here demonstrate how the tuning of the milling energy in the HEBM synthesis route of 1144 compounds has a significant influence on the product morphology. In the synthesis route, starting from the pure elements, a high milling energy is necessary to comminute some of the elements, such as the metallic calcium.

To be able to process powders through a milder HEBM treatment, we experimented with a novel approach for the synthesis of 1144 compounds, substituting bulk metallic calcium with calcium hydride powder. By opportunely tailoring the starting chemical composition and milling energy, significant refinement of grains can be obtained with respect to more energetic processes, obtaining powders characterized by a good degree of homogeneity and sharp superconducting transitions.