The synthesis and properties of LiFeAs, a high-T c Fe-based superconducting stoichiometric compound, are reported. Single crystal x-ray studies reveal that it crystallizes in the tetragonal PbFCl type ͑P4/nmm͒ with a = 3.7914͑7͒ Å and c = 6.364͑2͒ Å. Unlike the known isoelectronic undoped intrinsic FeAs compounds, LiFeAs does not show any spin-density wave behavior but exhibits superconductivity at ambient pressures without chemical doping. It exhibits a respectable transition temperature of T c = 18 K with electronlike carriers and a very high critical field, H c2 ͑0͒ Ͼ 80 T. LiFeAs appears to be the chemical equivalent of the infinite layered compound of the high-T c cuprates. DOI: 10.1103/PhysRevB.78.060505 PACS number͑s͒: 74.70.Dd, 61.66.Fn, 74.25.Fy Until recently the chemical realm of high-T c superconductivity had been limited mainly to copper oxide-based layered perovskites. The latest search for noncuprate superconductors in strongly correlated electron layered systems has led to the discovery of high-T c superconductivity in doped quaternary rare-earth iron oxypnictides, ROFePn ͑R = rare-earth metal and Pn= pnicogen͒. 1-3 These superconductors generated enormous interest in the materials community due to the high T c 's involved ͑up to 41-55 K͒ as well as the critical presence of a magnetic component, Fe, considered antithetical to conventional s-wave superconductivity. 3,4 High-pressure studies suggest maximum T c in R͑O,F͒FeAs may be about 50 K but higher T c 's ͑Ͼ50 K͒ may yet be discovered in structurally different compounds that are electronically related to R͑O,F͒FeAs. 5 Analogous alkaline-earth iron arsenides, AeFe 2 As 2 ͑Ae= Sr and Ba͒, reportedly having formal ͑Fe 2 As 2 ͒ 2− layers as in ROFFeAs but separated by simple Ae layers as in the cuprates, were found to behave similarly. 6,7 The AeFe 2 As 2 phases become superconducting ͑maximum T c ϳ 37 K͒ with appropriate substitution of Ae atoms with alkali metals. 8,9 It was also found that isostructural compounds KFe 2 As 2 and CsFe 2 As 2 with formal ͑Fe 2 As 2 ͒ 1− layers were superconducting, having much lower T c 's of 3.8 and 2.6 K, respectively. 9 Moreover, the evolution from a superconducting state to a spin-density wave ͑SDW͒ state by chemical substitution was observed in K 1−x Sr x Fe 2 As 2 . 9 Critical to the high-T c FeAs superconductors is the need to introduce sufficient amounts of charge carriers: with electrons ͑n type͒ by F doping ͑15-20 atm %͒ or holes ͑p type͒ by Sr doping ͑4-13 atm %͒ in ROFeAs, and ͑K/Sr͒ substitution ͑40: 60 atm %͒ in AeFe 2 As 2 . These results established the unique role of ͑Fe 2 As 2 ͒ layers in high-T c superconductivity. Since simple elemental K, Cs, ͑K/ Sr͒, or ͑Cs/Sr͒ layers separate the ͑Fe 2 As 2 ͒ layers in the AFe 2 As 2 superconductors, a Li-based analog, LiFeAs, was investigated. Its crystal structure was previously reported to be of the Cu 2 Sb type that features a Fe 2 As 2 substructure similar to the known FeAs superconductors. 10 However, the locations of the Li atoms were problematic....

Superconductivity and phase relationships were explored in the Na-Fe-As system. The PbFCl-type 111 phase is stable only within a Na stoichiometry range of 1.00 to ϳ0.85, and exhibits bulk superconductivity within an even narrower range around 0.90 in Na 0.9 FeAs. In particular, stoichiometric NaFeAs is not a bulk superconductor. The onset of the superconducting transition varies in a totally different way and the highest T c occurs in multiphase samples with a nominal composition of Na: Fe: As= 0.5:1:1, where the superconductive volume-fraction is almost zero. Such doping dependency is rather surprising and in disagreement with most expectations. DOI: 10.1103/PhysRevB.79.184516 PACS number͑s͒: 74.70.Dd, 74.62.Dh, 74.62.Bf The recent discovery of superconductivity in layered transition-metal oxypnictides, La͑O , F͒FeAs, 1 has attracted intense interest in the FeAs-based compounds. Superconductivity up to 55 K has been observed in three classes of FeAsbased compounds, i.e., ͑R ,Ae͒͑O , F͒FeAs, ͑Ae, A͒Fe 2 As 2 and AFeAs, where R, Ae, and A are rare earth, alkaline earth, and alkali elements, respectively.2-7 The FeAs-based superconducting compounds have often been compared with the well-investigated cuprate superconductors. The doping dependency of the superconductivity, however, appears to be rather different in the FeAs-family as it varies significantly from one member to another. 7,8 The main doping effects reported so far in the FeAs family, however, appear still to be a smooth, bell-like T c vs. carrier filling x 0 , where T c is the transition temperature. Competitions with magnetic ordering are often suggested in interpreting the data.9,10 Significant change in the superconducting volume-fraction V S , on the other hand, occurs only near the normal conductorsuperconductor boundary. The V S , it should be pointed out, is actually a convolution of the T c ͑x 0 ͒ and the local x 0 -distribution ͑composition inhomogeneity͒ if x 0 is a sole parameter. A constant V S , therefore, is expected if the superconductive range, ⌬, is much broader than the x 0 -spread, e.g., the full width at half height ͑FWHH͒ of a normal distribution. The effect on T c , in such cases, will be the main focus. At the opposite extreme of ⌬Ӷ, however, the spread would lead to the same T c distribution but a drastic V S change with x 0 , though this is rarely observed or discussed. Herein we report our observations in the superconducting system, Na y FeAs, which possesses a PbFCl-type structure isotypic to that of LiFeAs. This PbFCl-type structure as well as ͑trace͒ superconductivity exist over the whole nominalcomposition range investigated, i.e., with the nominal composition of Na y FeAs, with 0.5Յ y Յ 1.0. The samples are single phase, however, only for y Ն 0.9, and the impurity phase FeAs appears at lower y. A rather unusual doping effect is also observed. On one hand, the samples become bulk superconducting, e.g., with V S Ͼ 10%, only around y = 0.9 with an estimated spread Ӷ0.1. The apparent T c , on the other hand, monotonically...