Superconductivity in some heavy rare-earth iron arsenide REFeAsO1-δ (RE = Ho, Y, Dy and Tb) compounds

Here, we report some new iron arsenide superconductors of REFeAsO1−δ (RE = Ho, Y, Dy and Tb) that were successfully synthesized by a high-pressure synthesis method with a special rapid quenching process, with the onset superconducting critical temperatures at 50.3, 46.5, 52.2 and 48.5 K for RE = Ho, Y, Dy and Tb, respectively, as determined from the resistivity measurements.

2 superconducts at lower temperatures [8]- [14]. So far the electron pairing mechanism is still unclear in these superconducting materials, while all these iron arsenide superconductors have made up the second high-T c family after cuprates.
After the discovery of F-free REFeAsO 1−δ superconductors with RE = La, Ce, Pr, Nd and Sm [7], we have also investigated these quaternary compounds of all other heavy rare earth (RE) elements by a high-pressure (HP) synthesis method, and clear superconducting signals have been observed in some multi-phase samples with the T c around 50 K when RE = Gd, Tb, Dy, Ho and Y. However, a 1111 superconducting phase cannot be distinguished by x-ray diffraction (XRD) analysis except for RE = Gd [15]. It is speculated that the unobservable of 1111 phase by XRD was probably due to the line broadening effect of the tiny crystal grains and the shielding of abundant impurity phases. One possible explanation is that the 1111 phases with these heavy RE elements are metastable and easy to decompose. Independently, by synthesizing under a ultra-high pressure of about 12 GPa, F-doped 1111 phase of TbFeAs(O,F) and DyFeAs(O,F) were reported to be superconducting around 46 K [16].
Here, we report the high-T c superconductivity in some oxygen-deficient REFeAsO 1−δ (RE = Ho, Y, Dy and Tb) compounds, which were successfully synthesized and stabilized by a revised HP method with a rapid quenching approach.
The starting chemicals of Ho, Y, Dy and Tb chips and As, Fe and Fe 2 O 3 powders are all with a purity better than 99.99%. At the first step, REAs powder was obtained by reacting RE pieces and As powders at 650 • C for 12 h and 1180 • C for 40 h, and then thoroughly ground into a fine powder after cooling. For the HP synthesis, appropriate amounts of REAs, Fe and Fe 2 O 3 powders were mixed together according to a nominal chemical formula REFeAsO 0.8 and then ground thoroughly and pressed into small pellets. The pellet was sealed in boron nitride crucibles and then mounted on a six-anvil high-pressure synthesis apparatus. After applying a high pressure of 5 GPa, the temperature was increased to 1000 • C by electrical heating, and sintering for 1 h while keeping the pressure constant. Here, a special rapid cooling process was adopted. After switching off the power, the sample was quickly quenched to room temperature by water cooling within about 1 min and then the pressure was released (HP method). The rapid quenching process was supposed to keep the metastable phase that formed under high temperature and high pressure. The synthesized samples are dense and hard with black color. For comparison, we have also synthesized the above mentioned REFeAsO undoped compounds in a sealed evacuated quartz tube at 1180 • C for 60 h (AP method).
The phase purity and structural identification were characterized by powder XRD analysis on an MXP18A-HF-type diffractometer with Cu-K α radiation from 20 • to 80 • with a step of 0.01 • . The XRD results indicate that for the 1111 REFeAsO phase of the AP samples, only TbFeAsO compound was obtained, with the lattice parameters a = 3.898(1) Å and c = 8.404(2) Å, while other rare earth elements cannot form the parent 1111 phase by the AP method. The XRD pattern of the AP TbFeAsO sample was plotted in figure 1, together with a calculated pattern for identification of the 1111 phase. For the HP samples, besides the reported 1111 phases for RE = Dy and Tb, the same phases for RE = Ho and Y were successfully prepared here. The XRD patterns for HoFeAsO 1−δ and YFeAsO 1−δ were also plotted in figure 1, and the main diffraction peaks for the 1111 phase were pointed out by arrows. For all these samples with heavy RE metals, impurities always exist in the synthesized samples, and the main impurity phases were identified to be RE 2 O 3 , REAs and FeAs, which are all not superconducting in the measuring temperature, these impurities were identified in the case of YFeAsO 1−δ in figure 1. These REFeAsO 1−δ phases were speculated to be metastable and decomposed at a  lower temperature. To check this point, these HP samples were annealed in a vacuum quartz tube at various temperatures in the range of 600-1200 • C. All 1111 phases quickly disappeared with annealing except for RE = Tb (here, we note that for all O-deficient 1111 superconductors, the superconductivity would be destroyed by annealing, and the 1111 phase becomes undoped with the appearance of impurities). This explains why the quenching process worked to preserve the 1111 phase. The lattice parameters of all these synthesized 1111 phases were summarized in table 1. Clearly, with the increase of ionic size of the RE elements (Ho Y < Dy < Tb), the lattice parameters monotonically increase; and oxygen vacancy causes the decrease of lattice parameters as we have reported in [7]. The temperature dependence of resistivity for all samples was measured by a standard fourprobe method from 5 to 300 K. The AP samples were not superconducting while all HP samples exhibit superconducting zero resistance at low temperatures above 40 K (as shown in figure 2), with nearly linear resistivity versus temperature behavior above the superconducting transition,  The dc magnetization was measured using a Quantum Design MPMS XL-1 system. For an experimental cycle, the sample was cooled to 1.8 K in zero field cooling (ZFC) and data were gathered when warming in an applied field and then the sample was cooled under an applied field (FC) and data were collected again when warming up. The dc-magnetization data (measured under a magnetic field of 1 Oe) for all the HP samples of HoFeAsO 1−δ , YFeAsO 1−δ , DyFeAsO 1−δ and TbFeAsO 1−δ are shown in figure 3 with both the ZFC and FC results, with the onset diamagnetic transition temperature (determined from the ZFC curves) at 49.1, 45.2, 51.0 and 46.8 K, respectively. These are a little lower than the T c (onset) determined from the resistivity measurements, which can be attributed to the poorer quality of these superconductors than those with light RE elements [7], because of the difficulty in forming these heavy RE 1111 phases. Nevertheless, the diamagnetic superconducting shielding fractions for all these superconductors are from 30 to 75%, which indicates the bulk superconducting behavior of these compounds.
In conclusion, superconductivity has been achieved in some layered quaternary REFeAsO 1−δ compounds (for RE = Ho, Y, Dy and Tb) by the introduction of oxygen vacancy through HP synthesis, and the 1111 phase structure was successfully stabilized by a rapid quenching process. Therefore, together with previous discovery, high-T c superconductivity has been discovered in all these REFeAsO 1−δ compounds for RE = La, Ce, Pr, Nd, Sm, Gd, Tb, Dy, Ho and Y.