Surface layers and far infrared spectra of High-Tc superconductors

Abstract

It is possible to increase significantly the accuracy of far IR (FIR) transmission spectra of High-Tc superconductors by depositing a very thin oriented film (e.g. t = 300 Å), of either YBaCuO or GdBaCuO on a very thin MgO plate (e.g. t₀ = 90 μm). Such very thin films of superconductors are unstable and FIR transmission either at 300 or 10 K increases dramatically after a few weeks. A phenomenological model is however possible to explain all spectra at different temperatures from 300 to 10 K if we consider a thin layer of insulating barium cuprate (e.g. GdBaCuO₆) made by a progressive loss of oxygen.

More precisely, for one sample with T_c = 88 K, and within the accuracy of the measurements, the plasma frequency$\text{ν}{}_{\mathrm{p}}\text{=}\text{1}\text{2πc}\text{(}\text{N}{}_0\text{e}{}^2\text{mε}{}_0\text{ε}{}_{\mathrm{R}}\text{)}{}^{\mathrm{<mtext>1</mtext><mtext>2</mtext>}}$, is a constant,ν_p = 4500 cm⁻¹, and ν_c is a linear function of temperature for T >; Tc (ν_c = 1.6 T + 55 cm⁻¹). For T < Tc, a T⁴ law has been used to get the number of quasiparticles in the superconductive phase and to compute the spectrum at T = 86 K with the additional assumption of a collision frequency saturating around 50cm⁻¹. We have also to introduce 3 mid-IR active oscillators at room temperature, and to keep them in the superconductive phase, with a definite increase in their strengths.

That phenomenological model contains 11 parameters only since we have shown the possibility to use only one plasma frequency at all temperatures and to adjust only one collision frequency (the one at very low temperature).