Evidence that all high-temperature superconductors are p-type

Abstract

The ground charge state of Ce substituting for ${\mathrm{Nd}}^{+3}$ in ${\mathrm{Nd}}_{2}$ ${\mathrm{CuO}}_{4}$ is usually assumed to be ${\mathrm{Ce}}^{+4}$ , indicating that Ce is a donor which causes ${\mathrm{Nd}}_{2\mathrm{\ensuremath{-}}\mathrm{z}}$ ${\mathrm{Ce}}_{\mathrm{z}}$ ${\mathrm{CuO}}_{4}$ to be an n-type superconductor. Using the self-consistent bond/charge method, we show that the Ce ground state is ${\mathrm{Ce}}^{+3}$ , not ${\mathrm{Ce}}^{+4}$ , in ideal ${\mathrm{Nd}}_{2}$ ${\mathrm{CuO}}_{4}$ , implying that monatomic Ce dopants are isoelectronic centers, not n-type donors. The observed Ce doping of ${\mathrm{Nd}}_{2\mathrm{\ensuremath{-}}\mathrm{z}}$ ${\mathrm{Ce}}_{\mathrm{z}}$ ${\mathrm{CuO}}_{4}$ can be understood if an interstitial oxygen binds to the Ce, elevating its charge state to ${\mathrm{Ce}}^{+4}$ . This is a paired defect of Ce and oxygen, which dopes the host p-type, not n-type, because of oxygen's 2p holes. The condition that the interstitial oxygen does dope the host p-type translates into an approximate steric constraint that the ionic radius of the rare-earth ion ${\mathrm{R}}^{+3}$ must exceed that of ${\mathrm{Tb}}^{+3}$ for ${\mathrm{R}}_{2\mathrm{\ensuremath{-}}\mathrm{z}}$ ${\mathrm{Ce}}_{\mathrm{z}}$ ${\mathrm{CuO}}_{4}$ to superconduct. A similar situation occurs for the infinite-layer material ${\mathrm{Sr}}_{1\mathrm{\ensuremath{-}}\mathrm{v}}$ ${\mathrm{Nd}}_{\mathrm{v}}$ ${\mathrm{CuO}}_{2}$ , where the Nd dopant is isoelectronic ${\mathrm{Nd}}^{+2}$ , not donor ${\mathrm{Nd}}^{+3}$ as assumed, unless an extra oxygen is adjacent to the Nd. The data are reviewed in light of these results, and show no convincing evidence either of particle-hole symmetry in doping, or of n-type doping of any high-temperature superconductor. Instead they hint that the superconductivity is associated with oxygen, rather than with any particle-hole-symmetric cuprate-plane phenomenon.