vanderMarel

"Optics clues to pairing glues".

One possible interpretation of the infrared optical spectra of conducting materials uses the standard Kubo expression for a liquid of electrons coupled to a bosonic 'glue' described by a function $\alpha^2F(\omega)$ representing phonons, spin-fluctuations or other electronic degrees of freedom. We have applied this approach to a series of high T$_c$ superconducting materials, using an unbiased flexible function fitting of $\alpha^2F(\omega)$, with the aim to find out whether a consistent description of transport, optical and photo-electron spectra can be obtained. While it is probably justified to call into question the applicability of standard expressions for electron-boson coupling to these strongly correlated materials, the fitted $\alpha^2F(\omega)$ provides a compact and perhaps insightful way to characterize the transport and optical properties, which is deemed to merge with a conventional bosonic spectrum when the Fermi-liquid state is approached.

The outcome is, that the best fits of the optical data to a strong-coupling model are obtained with two components in the glue: one at 60 meV and the other at 250 meV. We observe qualitatively the same features in samples from underdoped to overdoped Bi2212, Bi2223, and Hg1201. The high energy scale of the fluctuations emanating from this analysis forms a peculiarity not traditionally associated with electron-phonon coupling, and implies strong coupling of the electrons among themselves, to spin-fluctuations or to other electronic degrees of freedom. While the overall frequency dependence of the thus-obtained 'glue' function appears to be significantly richer in structure than the marginal Fermi liquid model, one crucial feature of the latter model is also present in the experimental data: The 'glue' function calculated from the experimental data has significant temperature dependence at the low frequency end of the spectrum, a feature which is not normally associated with electron-phonon coupling.