Identification of prototypical Brinkman-Rice Mott physics in a class of iron chalcogenides superconductors

The 122$^{*}$ series of iron-chalcogenide superconductors, for example K$_x$Fe$_{2-y}$Se$_{2}$, only possesses electron Fermi pockets. Their distinctive electronic structure challenges the picture built upon iron pnictide superconductors, where both electron and hole Fermi pockets coexist. However, partly due to the intrinsic phase separation in this family of compounds, many aspects of their behavior remain elusive. In particular, the evolution of the 122$^{*}$ series of iron-chalcogenides with chemical substitution still lacks a microscopic and unified interpretation. Using angle-resolved photoemission spectroscopy, we studied a major fraction of 122$^{*}$ iron-chalcogenides, including the isovalently `doped' K$_x$Fe$_{2-y}$Se$_{2-z}$S$_z$, Rb$_x$Fe$_{2-y}$Se$_{2-z}$Te$_z$ and (Tl,K)$_x$Fe$_{2-y}$Se$_{2-z}$S$_z$. We found that the bandwidths of the low energy Fe \textit{3d} bands in these materials depend on doping; and more crucially, as the bandwidth decreases, the ground state evolves from a metal to a superconductor, and eventually to an insulator, yet the Fermi surface in the metallic phases is unaffected by the isovalent dopants. Moreover, the correlation-driven insulator found here with small band filling may be a novel insulating phase. Our study shows that almost all the known 122$^{*}$-series iron chalcogenides can be understood {\it via} one unifying phase diagram which implies that moderate correlation strength is beneficial for the superconductivity.
Comment: 19 pages, 5 figures