Common origin of superconductivity in hole- and electron-doped cuprates based on indirect-exchange coupling. Doping dependence of chemical potentials

The discovery in 1989 of superconductivity in “electron-doped” cuprates of composition Ln2−xCexCuO4−z: Ln=Nd, Pr, Sm, by Tokura et al. (Nature 337 (1989) 345) appeared at first to imply the existence of a novel branch of superconducting cuprates. Later, however, strong similarities with hole-doped cuprates were found at low temperatures such as positive Hall coefficients and striking analogies regarding (in-plane) resistivities. In the present paper we show, first of all, that also shifts in chemical potential upon doping, deduced from photoelectron spectra, are similar. They do not obey rules prescribed by a “semiconductor model” for either category of cuprates. Instead, the shifts appear to vary continuously in accordance with the value of a parameter in the (theta) function for conduction states, specifying the relative position of conduction and (main) valence bands. In a second step, we assert that these electron-doped cuprates are, in fact, hole-doped because of overcompensation of doped excess positive charge by oxygen anions. Taking account of established facts that the CuO2 layers in this (T′) structure are strained and locally heavily distorted in the superconductors we follow, in the framework of an indirect-exchange formalism for superconductivity earlier extensively applied for hole-doped cuprates, the development of critical temperature TC as a function of doping x. We reason that on this basis the superconducting doping range must be considerably less wide than in La2−xSrxCuO4 since the unstrained, undistorted (T′) structure does not afford superconductivity. In addition, analogies with TC(x) in the (La,Sr) cuprate for the Ce-underdoped, and with oxygen-over-doped Tl2Ba2CuO6+y in the Ce-overdoped range, must be expected. Detailed calculations of pressure gradients dTC/dP for the over(hole-)doped (Tl, Ba) compound strongly support this correspondence. In a recent paper by Fournier et al. [Phys. Rev. Lett. 81 (1998) 4720] both these analogies, with Pr2−xCexCuO4−z as the electron-doped cuprate, had been deduced from (in-plane) resistivities at low temperatures in high magnetic fields. The results of the present paper are also taken to imply that electron doping into insulating cuprates is not possible.