Synthesis and superconductivity in spark plasma sintered pristine and graphene-doped FeSe_0.5Te_0.5

1 Introduction

The discovery of layer structured cuprates, with superconducting transition temperatures (T_c) above the liquid-nitrogen temperature, has heralded the search for high T_c superconductors which transcend the traditional Bardeen-Cooper-Schrieffer type superconductors. In cuprates, superconductivity evolves from doping of antiferromagnetic Mott insulating phase with excess carriers. Following this discovery, many materials have been found to exhibit superconductivity when their intrinsic long-range magnetic order is suppressed by doping beyond a threshold dopant concentration. In this regard, the recent realization of five different Fe-pnictide superconductors [1–24] in LaFeAsO_1-xF_x, BaFe₂As₂, LiFeAs, Sr₂PO₃FePn, and FeSe_xTe_1-x systems has reinvigorated the search for new high-T_c non-cuprate materials. While a comprehensive understanding of superconductivity in these systems is still under development, it has been proposed (from an experimental standpoint) that some magnetic fluctuations (e.g. Fe 3d orbital fluctuations across the Fermi energy) may play a similar role as phonons for the formation of Cooper pairs [5]. These unusual electron pairing mechanisms in Fe-based materials, therefore, offer novel insights onto the long-standing fundamental question on the pair mechanism of high-T_c superconductors, cuprates and non-cuprates. Among the five different recently discovered Fe-based superconductors [1–24], FeSe is the only pnictide with a simple crystal structure [1, 3, 6], and is therefore highly suitable for elucidating the fundamental mechanism of superconductivity. Consequently, the crystal structure of FeSe may be easily distorted by external pressure or doping for an increased T_c. Although many researchers have reported on the influence of doping and external pressure in FeSe materials [6], there are still two major questions that warrant a focused study: (i) from a fundamental perspective, a coherent understanding of the interplay between magnetism and superconductivity still remains to be achieved [1, 3, 6], and (ii) the low critical current density (J_c) of FeSe materials is a persisting concern for practical applications [1]. Since J_c is intimately related to the micro-morphology of grain boundary, it is highly desired to synthesize FeSe materials with controlled grain boundary morphology [1].

In the present work, we address these challenges by employing a novel synthesis technique through ball-milling and subsequent spark plasma sintering (SPS) without any need for pre-alloying. We expect that the SPS method results in lower magnetic impurities and also facilitates the introduction of defects/pinning centers (for example, through the addition of nanomaterials such as graphene). SPS is a high energy, low voltage, pulsed plasma discharge in a low-pressure atmosphere that can generate highly localized Joule heating (up to a few thousand degrees Kelvin) in a few minutes [1, 25, 26]. Here, we employed in situ SPS to synthesize the superconducting phase of FeSe_0.5Te_0.5 to simultaneously enhance both the T_c and J_c. Moreover, since FeSe_0.5Te_0.5 is a layered material, we included few-layer graphene during the ball-milling and SPS process for creating vortex pinning centers, and exploring the interplay between magnetism and superconductivity. We observed that our samples exhibited a co-existence of ferromagnetic ordering and superconductivity below T_c=15 K. Notably, our magneto-resistance data showed no loss in superconductivity up to 7 T with a <800 mK shift in T_c. We explain our results in terms of a self-consistent superconducting shell-magnetic core scheme wherein the shell forms a percolated robust superconducting path.

2 Materials and methods

The starting materials for the synthesis of pristine FeSe and FeSe_0.5Te_0.5 were as-purchased powders of Fe (~100 μm, >99.9%), Se (~50 μm, >99.9%), Te (~50 μm, >99.9%) from Alfa-Aesar (Ward Hill, MA, USA). These elemental powders were mixed according to the chemical formula to yield FeSe_0.5Te_0.5 in an Ar purged glove box. Subsequently, the powders were ball-milled in a stainless steel container for 15 h under Ar (manufacturer: SPEX sampleprep LLC., Metuchen, NJ, USA) with 1:3 power-to-ball ratio. The ball-milled powders were sintered using the SPS technique at 700°C for 5 min at a pressure of 50 mPa with ON-OFF ratios of 12:2 and 1:1 (Fuji electronics Co. Ltd., Dr. Sinter, Kanagawa, Japan). Few-layer graphene sheets were prepared using a chemical exfoliation technique described in Ref. [27].

3 Results

As shown in Figure 1, the examination of fractured surface for sintered specimens revealed a poly-crystalline and layered structure for pristine FeSe_0.5Te_0.5 samples with layers extending to tens of microns. Furthermore, the presence of graphene (0.5–2% wt.) appeared to increase the density (from ~85% to 90% of theoretical density) of sintered samples, indicating a possible lattice distortion. Unlike many previous studies, we used the SPS process to directly synthesize the superconducting phase rather than using it as a mere densification tool. The advantage of this process lies in its high throughput and short timescales (~5–10 min) required for sample preparation without the need for any pre-alloying. The X-ray diffractograms (Supplemental Figure 1) showed that the observed interplanar distance matches with FeSe_0.5Te_0.5 in accordance with Vegard’s law.

Figure 1:

Representative scanning electron micrographs for FeSe_0.5Te_0.5 samples with: (A) 0% wt., (B) 0.5% wt., (C) 0.8% wt., and (D) 2% wt. graphene.

All SPS synthesized samples exhibited a T_c with a clear signature for ferromagnetism (see Figure 2) both below and above T_c (~15 K), indicating the co-existence of FM and superconducting domains. The evident drop in resistivity ~15 K (Figure 3) and a simultaneous display of diamagnetic behavior (as shown in Figure 2) confirms the presence of a percolating superconducting path among the FM domains. Interestingly, the magnetic hysteresis did not exhibit much temperature dependence below 15 K. Such an observation may plausibly be attributed to field trapping in FM domains during M-H sweeps above T_c. The magnetic properties of the FeSe family of compounds are extremely sensitive to the crystal structure, stoichiometry and the presence of excess Fe. For instance, superconductivity in Fe_1.01Se with T_c~8 K is destroyed due to FM ordering introduced by only 0.2 at% increase in Fe, and no superconductivity is observed in Fe_1.03Se [28]. In this regard, it is possible that any unreacted Fe or formation of Fe_1-ySe_0.5Te_0.5 phase during the SPS process may contribute to the observed FM signature (Figure 2). Indeed, the nominal composition of our samples lies in the FM ordering and superconductivity overlapping region of the phase diagram of Fe_1.03Se_0.5Te_0.5 described in Ref. [9]. The normalized magnetic moment (M_N=M[5 K]-M[20 K]) clearly exhibits a shape typical of superconductors showing evident changes at H_c1 and H_c2 for all samples (Figure 2). While the magnetic data suggests an H^M_c2 (i.e. H_c2 from magnetic measurements) ~8–10 kOe, the resistivity (ρ) data exhibited a clear superconductivity transition ~14.5 K in external magnetic fields as high as 7 T (Figure 3). Interestingly, H^M_c2 was found to be highest at 0.5% wt. of graphene doping (performed during ball-milling) suggesting that the magnetic vortices can be pinned effectively by graphene at low concentrations <1% wt. (Figure 4). As shown in Figure 3, the superconducting transition persisted even at external fields as high as 7 T in ρ(T) data suggesting that the H^R_c2 [i.e. H_c2 from ρ(T) data measurements] is much higher than H^M_c2, as observed in other Fe-based superconductive systems. Such an observation may plausibly be attributed to the presence of magnetic domains that result in a pseudo H_c2 value in M-H measurements shown in Figure 2. As shown in the insets of Figure 3A–D, we determined T_c values by identifying the peak position of dρ(T)/dT. We observed that graphene dopants increased T_c of FeSe_0.5Te_0.5 samples at low concentrations (<1% wt.) by 200 mK (Figure 4). The net shift in T_c (ΔT_c=T_c[0 T]-T_c[7 T]) upon the application of external field was found to be within 400–800 mK suggesting a robust superconducting percolated transport path in our samples.

Figure 2:

Hysteresis loops showing magnetization vs. the applied field for all the FeSe_0.5Te_0.5 samples with (A) 0% wt., (B) 0.5% wt., (C) 0.8% wt., and (D) 2% wt. graphene. The normalized magnetic moment (M_N=M[5 K]-M[20 K]) clearly exhibits a shape typical of superconductors showing evident changes at H_c1 and H_c2 for all samples.

Figure 3:

Temperature dependent electrical resistivity for all the FeSe_0.5Te_0.5 samples with (A) 0% wt., (B) 0.5% wt., (C) 0.8% wt. and (D) 2% wt. graphene. A transition is observed ~14.5 K suggesting that there is a superconducting percolation path in the sample. The insets show the derivative of resistance with respect to temperature and help in clearly identifying T_c. For all pristine and doped FeSe_0.5Te_0.5 samples, superconductivity (SC) transition was observed up to 7 T suggesting robust superconductivity.

Figure 4:

The effect of graphene doping on critical fields H_c1 and H_c2 (obtained from magnetism measurements) and T_c obtained from the plots shown in Figures 2 and 3.

The pulsed direct current (DC) method of energizing in the SPS method activates particulate surfaces by removing either impurities or oxidation layers through electric discharges between neighboring particulates in addition to Joule heating at the grain boundaries. In the SPS process, the applied DC pulses are square/rectangular waveforms with different ON-OFF durations. The ON time activates the particulates while the OFF time allows relaxation of the sintered material during the sintering process. Hence, the ON-OFF time ratio of x:y suggests that there are x ON current pulses with y OFF pulses used during the sintering process. The time duration of each ON or OFF pulse is 2.7 ms (for further details, see SPS standard Procedure, Dr. Sinter SPS, Service Manual No. 16). It is well-known that a change in SPS ON-OFF conditions can significantly influence material properties such as electrical resistivity, mechanical robustness and thermopower. In this regard, to study the effect of SPS pulse conditions, we prepared FeSe_0.5Te_0.5 samples at 12:2 and 1:1 ON-OFF ratios. Interestingly, we did not observe any differences in the magnetic and electrical transport between these samples, suggesting that the superconductivity in our samples is robust.

Juxtaposing the data from magnetic, transport and SPS ON-OFF ratios, we propose a superconducting shell-FM core structure (see Figure 5) with a percolated conductive path among the superconducting shells. Since Fe is more conducting than Se and Te, it is expected that the DC passes mainly through Fe powders. As shown in Figure 5, such a condition results in the formation of a FeSe_0.5Te_0.5 shell with some unreacted or partly reacted Fe, Se, and Te forming a FM core. While the cores exhibit traditional FM ordering above T_c, any trapped field can result in a remnant magnetization below T_c despite the shielding of FM cores by superconducting shells. In the case of low graphene dopant concentrations, graphene is possibly present at the interfaces between the FM core and superconducting shells acting as a pinning center that prevents the vortices from spreading into the superconducting path.

Figure 5:

A schematic of the FeSe_0.5Te_0.5 sample formed through ball-milling and spark plasma sintering (SPS).

4 Conclusion

In summary, we present a new ball-milling and SPS-based technique for the facile synthesis of FeSe_0.5Te_0.5 samples without any pre-alloying. We observed that FeSe_0.5Te_0.5 exhibits a coexistence of FM ordering and superconductivity signature plausibly arising from a FM core-superconductivity shell structure. We doped the samples with graphene during the synthesis process for vortex pinning. We found that graphene could be doped up to 2% wt. using ball-milling and the SPS method, but did not show any significant effects on the T_c. Interestingly, 0.5% wt. of graphene resulted in a slight increase (~400 mK) in T_c while doping above 0.5% wt. deteriorated the superconductive properties. For all the doping ratios, we observed that the superconducting transition in electrical transport measurements is robust up to 7 T.