Effects of annealing, acid and alcoholic beverages on Fe_1+yTe_0.6Se_0.4

The recent discovery of superconductivity at 26 K in an iron oxypnictide, LaFeAs(O, F) [1], has stimulated great interest among condensed-matter physicists. A tremendous amount of work has been carried out, leading to the emergence of novel iron-based superconductor families with different crystal structures: 1111 (LaFeAs(O, F)) [1], 122 ((Ba,K)Fe₂As₂) [2], 111 (LiFeAs) [3] and 11 (Fe(Te, Se)) [4]. Among these, FeTe_1−xSe_x has received special attention due to its simple crystal structure, comprising only Fe(Te, Se) layers. Superconductivity and magnetism of this system are not only dependent on the doping level but are also sensitive to Fe non-stoichiometry, which originates from the partial occupation of excess Fe at the interstitial site in the Te/Se layer [5, 6]. For the undoped parent compound Fe_1+yTe, the commensurate antiferromagnetic order can be tuned by the excess Fe to an incommensurate magnetic structure [5]. In Se-doped FeTe samples, excess Fe was found to suppress superconductivity and cause the magnetic correlations [6]. In the other end member, Fe_1+ySe, superconductivity is reported to reside just in a very narrow concentration region of excess Fe [7]. To probe the intrinsic properties of FeTe_1−xSe_x without the influence of excess Fe, some previous works have been performed to remove the effect of excess Fe by annealing it in different conditions. Superconductivity was found to be induced by annealing in a vacuum [8, 9], air [10], oxygen [11, 12], nitrogen [13], I₂ vapour [14] and even by immersing into nitric acid HNO₃ [13]. Recently superconductivity was also reported to be induced in FeTe_0.8S_0.2 by immersing it into alcoholic beverages [15, 16]. The interpretations for these treatments for inducing superconductivity are still controversial, including the improvement of homogeneity [8], oxygen intercalation [11] and deintercalation of excess Fe [10, 13, 14]. Although superconductivity can be successfully induced, some annealed samples show a very broad transition width [11, 13, 14, 16], which implies an inhomogeneous superconductivity. In order to address these issues, and try to understand how the superconductivity is induced in Fe_1+yTe_0.6Se_0.4, we have systematically investigated different methods to induce superconductivity by annealing in a vacuum, N₂, O₂ and I₂ atmospheres, and immersing the samples into acid and alcoholic beverages. By comparing the results, we give a possible scenario to explain how the superconductivity is induced in the non-superconducting as-grown Fe_1+yTe_0.6Se_0.4.

Single crystals with nominal compositions FeTe_0.6Se_0.4 and FeTe_0.8S_0.2 were prepared from high purity Fe (99.99%), Te (99.999%), Se (99.999%) and S (99.9999%) grains [8]. More than 10 g of stoichiometric quantities was loaded into a small quartz tube with d₁ ∼ 10 mm ∅ and the tube evacuated and sealed. Then we sealed this tube into a second evacuated quartz tube with d₂ ∼ 20 mm ∅. The whole assembly was heated up to 1070 °C and kept for 36 h, followed by slow cooling down to 710 °C at a rate of 6 °C h⁻¹. After that, the temperature was cooled down to room temperature by shutting down the furnace. The as-grown single crystal with typical dimensions of 5.0 × 5.0 × 1.0 mm³ for FeTe_0.6Se_0.4 and 2.0 × 2.0 × 0.1 mm³ for FeTe_0.8S_0.2 can be easily cleaved perpendicular to the c axis. For vacuum annealing, the sample was loaded into a quartz tube, which was carefully baked and examined to ensure that no appreciable amount of gas was emitted under the same conditions as the sample annealing. The quartz tube was carefully evacuated by a diffusion pump before sealing the sample. For annealing with I₂, the sample was loaded into a quartz tube with I₂ chips (99.9%). For annealing in O₂ and N₂ atmospheres, after evacuating, we filled controlled amounts of O₂ and N₂ gas before sealing the quartz tube. During these processes, a diaphragm-type manometer with an accuracy of 10⁻³ Torr was used for real-time monitoring of the pressure in the system to prevent gas leakage and to check the amount of gas in the tube. Then the samples were annealed at 400 °C for more than a day, followed by quenching in water. For O₂ annealing, we tried annealing the sample under 0.1%, 1% and 1 atmosphere (atm) of O₂ gas. Fe_1+yTe_0.6Se_0.4 samples annealed both under 0.1% and 1% atm of O₂ show a superconducting transition temperature, T_c, higher than 14 K, and the former exhibits larger J_c. By contrast, the sample annealed in 1 atm O₂ was totally damaged, which indicates that the sample itself was oxidized during annealing in too much O₂. Other pieces of as-grown samples were put into glass bottles (10 ml) filled with 20% hydrochloric acid HCl, beer (Asahi Breweries Ltd), red wine (Asahi Breweries Ltd), Japanese sake (Hakutsure Sake Brewing, Co. Ltd), shochu (Iwagawa Jozo Co. Ltd) or whisky (Suntory Holdings Ltd). The samples immersed in alcoholic beverages (beer, red wine, Japanese sake, shochu and whisky) were kept at 70 °C for 40 h. The sample immersed into 20% HCl was kept at room temperature for 100 h, because we found heating the sample up to 70 °C with acid damaged the sample quickly. Magnetization measurements on the as-grown and post-treated samples were performed using a commercial superconducting quantum interference device magnetometer (MPMS-XL5, Quantum Design). Magneto-optical (MO) images were obtained by using the local field dependent Faraday effect in the in-plane magnetized garnet indicator film employing a differential method [17, 18].

Figure 1(a) shows the temperature dependence of zero-field-cooled (ZFC) and field-cooled (FC) magnetization at 5 Oe for the as-grown, vacuum, N₂, O₂ and I₂ atmosphere annealed Fe_1+yTe_0.6Se_0.4 single crystals. The as-grown crystals usually show no superconductivity or a very weak diamagnetic signal below 3 K. These results show that superconductivity cannot be induced by vacuum or N₂ atmosphere annealing, which is quite different from previous reports [8, 9, 13]. Annealing in an I₂ atmosphere can successfully induce superconductivity, while the T_c is lower and the transition width is much broader than that in the crystal annealed in 0.1% O₂, which shows a T_c higher than 14 K with a transition width less than 1 K (obtained from the criteria of 10 and 90% of the magnetization result at 2 K). We also tested annealing crystals in a poor vacuum (∼1 Torr), which also enhanced the superconductivity. Thus, the previously reported vacuum or N₂ atmosphere annealing induced superconductivity is probably due to the poor vacuum or air leakage during the annealing process [8, 9, 13]. Figure 1(b) shows the temperature dependent magnetization of Fe_1+yTe_0.6Se_0.4 crystals immersed into alcoholic beverages and 20% HCl, together with the result of O₂ annealing for comparison. All these crystals exhibit superconductivity, among which the crystal immersed into beer has the largest diamagnetic signal. Although superconductivity can be induced by alcoholic beverages and acid, similar to the crystal annealed with I₂, T_c is lower and the transition width is much broader than that annealed in O₂ atmosphere. Alcoholic beverage induced superconductivity was also reported in a S-doped FeTe sample [16]. For comparison, temperature dependent magnetization of Fe_1+yTe_0.8S_0.2 single crystal immersed into beer at 70 °C for 1 week is also plotted in figure 1(b), which shows a T_c of about 7 K, close to the previous report [16]. In this case, the transition width is also very broad, similar to the effect of alcoholic beverage on Fe_1+yTe_0.6Se_0.4 crystals, and the diamagnetic signal is even smaller than that in Fe_1+yTe_0.6Se_0.4 treated by all of the methods.

Zoom In Zoom Out Reset image size

It is well known that even if the sample is mostly non-superconducting, the diamagnetic signal becomes significant when the non-superconducting region is covered by a superconducting region. Thus, to further confirm the nature of superconductivity, magnetic hysteresis loops (MHLs) of samples immersed into alcoholic beverages, acid as well as annealed in an I₂ atmosphere were measured at 2 K and are shown in figure 2(a). Although an obvious superconducting hysteresis loop can be seen, it is too weak to persist up to applied fields larger than 20 kOe. The absolute value of magnetization is also very small compared with that of O₂ annealed Fe_1+yTe_0.6Se_0.4 crystals, as shown in figure 2(b).

Zoom In Zoom Out Reset image size

From the MHLs, we can obtain the critical current density, J_c, by using the Bean model [19]:

$$\begin{eqnarray} &&{J}_{\mathrm{c}}=2 0\frac{\Delta M}{a(1-a/3 b)}, \end{eqnarray} \tag{ 1 }$$

where ΔM is M_down–M_up, M_up and M_down are the magnetization when sweeping fields up and down, respectively, a and b are sample widths (a < b). Magnetic field dependence of J_c at 2 K is summarized in figure 3. J_c of the O₂ annealed sample reaches a value greater than 5 × 10⁵ A cm⁻² at zero field. It is robust under applied field, keeping a value larger than 1 × 10⁵ A cm⁻² even at 50 kOe. Although the value of J_c is still lower than that of Ba(Fe_1−xCo_x)₂As₂ single crystal [20–22], it is the largest among those reported in Fe(Te, Se) [8, 23–28]. This fact demonstrates the high quality of our O₂ annealed sample. To further confirm the homogeneous current flow within the sample, we took MO images of 0.1% atm O₂ annealed crystal in the remanent state. In the inset of figure 3, a typical MO image taken at 6 K after cycling the field up to 800 Oe along the c axis is shown. The MO image manifests a typical roof-top pattern, similar to that observed in high quality Ba(Fe_1−xCo_x)₂As₂ single crystals [20, 29, 30], indicating a nearly uniform current flow in the crystal. The uniformity of J_c is much better than that reported in [8]. The profile of the magnetic induction along the dashed line in the MO image is also shown in the inset of figure 3. J_c can be roughly estimated by J_c ∼ ΔB/t, where ΔB is the trapped field and t is the thickness of the sample. With ΔB ∼ 1076 G and t = 50 μm, J_c is estimated as 2.2 × 10⁵ A cm⁻² at 6 K, consistent with estimation from the Bean model. The large J_c estimated from MHLs and the MO image cannot be sustained only by superconductivity near the surface. Thus, the superconductivity induced by O₂ annealing must be a bulk property of the sample. In Fe_1+yTe_0.6Se_0.4 single crystal immersed in alcoholic beverages or acid as well as annealing with I₂,J_c at 2 K under zero field is smaller than 1×10⁴ A cm⁻². In addition, J_c decreases very quickly with increasing field, and becomes smaller than 1 × 10² A cm⁻² at fields above 20 kOe. For FeTe_0.8S_0.2, the value of the zero field J_c is smaller than 500 A cm⁻² and easily suppressed by the modest applied field, which decreases to less than 100 A cm⁻² at fields about 10 kOe. The much smaller value of J_c compared with that in the O₂ annealed sample shows that the superconductivity comes from the surface.

Zoom In Zoom Out Reset image size

To directly prove that the superconductivity induced by I₂, acid and alcoholic beverages is only near the surface, we cut the four edges of the sample, polished the double surfaces (more than half of the sample was removed after polishing), and then compared the superconducting property of the whole sample and the inner part. In figures 4(a) and (b), the results on samples annealed in 1% atm O₂ and immersed into beer are compared. In the case of the O₂ annealed sample, T_c of the inner part is almost the same as that in the whole sample, and the diamagnetic signal as well as J_c do not decrease but become a little larger than the whole sample. It should be noted that the 1% atm O₂ annealed sample shows smaller J_c than that annealed under 0.1% atm O₂, although T_c is similar. The decrease in J_c may be caused by the fact that the sample itself was partially oxidized by the excess O₂, especially on the surface, which changes colour after annealing, as reported before [10]. After removing the over-oxidized surface, the diamagnetic signal and J_c are increased. By contrast, in the sample immersed in beer, both the diamagnetic signal and the J_c of the inner part are decreased to about 10% of the whole sample. The significant degradation in superconducting properties after removing the surface directly proves that the observed superconductivity mainly comes from the surface.

Zoom In Zoom Out Reset image size

Compiling the obtained results, we propose a scenario to explain how the superconductivity is induced by different methods. Vacuum and N₂ annealing cannot induce superconductivity, which indicates that homogeneity is not the key factor for inducing superconductivity. The previously reported vacuum and N₂ annealing induced superconductivity may be caused by the leakage of O₂, since we found that only trace amount of O₂ can successfully enhance superconductivity. Furthermore all oxidizing agents, O₂, I₂, acid and alcoholic beverages can induce superconductivity, possibly by removing the excess Fe in the sample rather than the O₂ itself being doped into the crystal. Among these, only O₂ can easily intercalate between the layers of the single crystal, and deintercalates the excess Fe from the inner part of the sample because of its small size, inducing bulk superconductivity in Fe_1+yTe_1−xSe_x. On the other hand, I₂, acid and the alcoholic beverages can just induce superconductivity near the surface because their relatively larger size prevents them from intercalating between the layers of the crystal. Thus, the rate of excess Fe deintercalation is much slower than the O₂ annealing, and can induce superconductivity only near the surface of the crystal. Very recently, the alcoholic beverage effect was also reported in Fe_1+yTe_0.9Se_0.1 polycrystal [31].

O₂ annealing and alcoholic beverages have been reported to be effective in Fe_1+yTe_1−xS_x single crystals [15, 32]. While for O₂ annealed Fe_1+yTe_1−xS_x single crystals, superconductivity can be suppressed by the following vacuum annealing, which is quite different from the case of O₂ annealed Fe_1+yTe_1−xSe_x, in which the superconductivity is stable under vacuum annealing (data not shown). This suggests that the oxygen is doped into Fe_1+yTe_1−xS_x rather than deintercalating the excess Fe. However, although superconductivity is successfully induced in Fe_1+yTe_1−xS_x, there is still no proof for bulk superconductivity like large J_c or an obvious jump of specific heat at T_c. More efforts along these lines are necessary to clarify the origin of different effects between Fe_1+yTe_1−xSe_x and Fe_1+yTe_1−xS_x single crystals.

In conclusion, we have found that vacuum and N₂ annealing cannot enhance superconductivity in Fe_1+yTe_0.6Se_0.4 single crystal. O₂ and I₂ annealing, acid and alcoholic beverages can induce superconductivity by oxidizing the excess Fe in the sample. The large value of J_c ∼ 5 × 10⁵ A cm⁻² obtained from MHLs and the MO image shows that bulk superconductivity can be induced by O₂ annealing. The self-field J_c at 2 K of the samples annealed in an I₂ atmosphere or immersed in acid or alcoholic beverages are smaller than 1 × 10⁴ A cm⁻², and suppressed by the modest magnetic field. Furthermore, the diamagnetic signal and J_c of the inner part of these samples are much smaller than those obtained from the whole sample. These results indicate that the I₂ atmosphere, acid and alcoholic beverages can only induce superconductivity near the surface.

This work is partly supported by the Natural Science Foundation of China, the Ministry of Science and Technology of China (973 project no. 2011CBA00105).