Hydrogen in layered iron arsenides: Indirect electron doping to induce superconductivity

Utilizing the high stability of calcium and rare-earth hydrides, CaFeAsF${}_{1-x}$H${}_x$ ($x$ $=$ 0.0–1.0) and SmFeAsO${}_{1-x}$H${}_x$ ($x$ $=$ 0.0–0.47) have been synthesized using high pressure to form hydrogen-substituted 1111-type iron-arsenide superconductors. Neutron diffraction and density functional calculations have demonstrated that the hydrogens are incorporated as H${}^{-}$ ions occupying F${}^{-}$ sites in the blocking layer of CaFeAsF. The resulting CaFeAsF${}_{1-x}$H${}_x$ is nonsuperconducting, whereas, SmFeAsO${}_{1-x}$H${}_x$ is a superconductor, with an optimal $T_{\mathrm{c}}$ $=$ 55 K at $x$ ∼ 0.2. It was found that up to 40% of the O${}^{2-}$ ions can be replaced by H${}^{-}$ ions, with electrons being supplied into the FeAs layer to maintain neutrality (O${}^{2-}$ $=$ H${}^{-}$ $+$ $e^{-}$). When $x$ exceeded 0.2, $T_{\mathrm{c}}$ was reduced corresponding to an electron overdoped region.

Figure 1

(a) Powder XRD patterns of CaFeAsH and CaFeAsD at room temperature (RT). (b) TG-MS (m/z $=$ 2 corresponds to the H${}_2$ molecule) profiles of CaFeAsH. Weight loss due the decomposition of the sample with hydrogen emission was observed from 200 °C to 600 °C. Hydrogen concentration was estimated to be 1.05 molecules per unit cell by integration of the MS trace curve. (c) NPD pattern of CaFeAsD observed at 300 and 10 K. (d) Crystal structure of CaFeAsH at 300 K (tetragonal) and 10 K (orthorhombic).

Figure 2

(a) ρ-$T$ profile of CaFeAsH and CaFeAsD compared with CaFeAsF. (b) Total DOS and atomic projected density of states (PDOS) of CaFeAsH obtained by density functional calculation using the VASP code. The origin was set at the Fermi level. (c) Variation of lattice constants ($a$, $c$) as a function of $x$ in the solid solution CaFeAsF${}_{1-x}$H${}_x$. Hydrogen content ($x$) was estimated by the TG-MS method.

Figure 3

Structural and chemical composition data on SmFeAsO${}_{1-x}$H${}_x$. (a) Powder XRD patterns of SmFeAsO${}_{1-x}$H${}_x$ (nominal $x$ $=$ 0.3, 0.4, and 0.5). Insets show close-up views. (b) TG-MS (m/z $=$ 2) profiles with nominal $x$ $=$ 0.15. H${}_2$ gas emission associated with weight loss was observed from 400 °C to 800 °C. (c) Oxygen deficiency content determined by EPMA measurement ($y$) and hydrogen content estimated by the TG-MS method ($x$) in SmFeAsO${}_{1-y}$H${}_x$ as a function of nominal $x$ in the starting mixture. The measured $x$ is almost equal to $y$ and nominal $x$, indicating the deficiency of the oxygen site is wholly compensated by the occupation of hydrogen. (d) Lattice constants ($a$ and $c$) as a function of $x$.

Figure 4

Electrical and magnetic properties. (a) ρ-$T$ profiles of SmFeAsO${}_{1-x}$H${}_x$ in underdoped (left, $x$ $=$ 0.0–0.19) and overdoped states (right, $x$ $=$ 0.22–0.47). (b) Log [ρ($T$) − ${\rho }_0$] vs log $T$ plots for data on the nonsuperconducting region, where ${\rho }_0$ is the residual component of resistivity obtained from fitting (see text). Lower parts show ${\rho }_0$ and the exponent residual resistivity $n$ as a function of $x$. (c) ZFC and FC 4πχ-$T$ curves measured under the magnetic field ($H$) of 10 Oe. (d) Magnetization ($M$) vs $H$ curves and shielding volume fraction vs $x$ plot at 2 K. (e) $x$-$T$ diagram of SmFeAsO${}_{1-x}$H${}_x$ superimposed by that of SmFeAsO${}_{1-x}$F${}_x$.[33]