Fresh off the ArXiv: Mysteriously high superconducting transition temperature in FeSe/STO

Fresh off the ArXiv: Mysteriously high superconducting transition temperature in FeSe/STO

Three FeSe/STO sample configurations

Bulk FeSe has a superconducting transition of about 8 K, low by even iron-based superconductor standards. So you can imagine the surprise of researchers at the discovery of ˜100 K superconductivity in single layer FeSe on SrTiO3 (STO). The highest superconducting transition temperature, Tc, of any iron-based material tops out a little above 50 K and FeSe can only reach 36 K under pressure.

The FeSe/STO case is even more tantalizing because it’s likely that interface effects are important here. Condensed matter folks just love interfaces. They’re a perfect breeding ground for new phenomena: proximity effects, reduced symmetry, doping and strain influences all come into play.

FeSe/STO crystal structure
FeSe/STO crystal structure. Image source.

The origin of this anomalously high Tc does make for an intriguing mystery in the “interface genre”. Pairing in bulk FeSe is thought to be unconventional (that is, non-phononic). The mechanism likely involves spin fluctuations mediating pairing of carriers between electron-like and hole-like Fermi surfaces. However, in the FeSe/STO case, the hole pocket is pushed down below the Fermi level, leaving only two electron bands behind, near the Brillouin zone boundary. Can the same mechanism still be at play?

Mechanisms to enhance superconductivity involving interface doping or tensile strain have been put forward and largely excluded. To add another suspicious character to the scene, ARPES measurements show evidence of shadow bands that look to be related to interface oxygen optical phonons, polarized out of the STO plane. Could this novel superconductivity be phonon-mediated after all?

Or perhaps electronic correlations are enhanced by electron-phonon interactions? Or, as others argue, would the electron-phonon interaction be entirely screened, playing no role at all?

A red herring in our midst.

Into the fray come two new papers fresh off the arXiv, only five days apart. The first uses isotope effects to tune the interface phonons and concludes that electron-phonon interactions are crucial for the enhanced superconductivity. The second uses scanning tunneling microscopy (STM) and electron energy loss spectroscopy (EELS) to study the interfacial phonons, and comes to the opposite conclusion: the phonons play no role.

Isn’t science fun?

Let’s break down their main findings.

No, phonons do not drive enhanced superconductivity

Let’s start with the argument against. This study was led by collaborators in Karlsruhe, Germany.

Electron Energy Loss Spectroscopy (EELS) measurements found two interfacial STO phonons at 57 and 91.5 meV. The 91.5 meV phonon is the one thought to be responsible for the Tc increase, and for the shadow band seen in ARPES. However the authors find the intrinsic linewidth of this phonon is nearly q-independent, contrary to what one would expect for strong electron-phonon coupling at zone center (q=0), and for this phonon to be the origin of the shadow band.

Interfacial phonons of FeSe/STO
Two interfacial phonons of FeSe/STO. Figure source.

Tunneling spectra measured below Tc are consistent with the presence of an excitonic bound state expected for the electronic pairing mechanism. The authors recorded an atomically-resolved map of the FeSe single layer. The 57 meV phonon is seen, but only near a grain boundary where the lattice is shifted by a half unit cell. However the superconducting gap does not change significantly away from this boundary.

The authors argue that this observation, and the fact that the 91.5 meV phonon is not observed, support the argument that the interfacial phonon cannot be responsible for the increased Tc. They suggest the observation of the 91.5 meV phonon in EELS and ARPES measurements may simply be due to phonon excitation by the electric field of the free photo-electron and its induced mirror charge.

A red herring after all.

Yes, phonons DO drive enhanced superconductivity!

Ah but not so fast. A group of collaborators in China, led by researchers at Fudan University in Shanghai, have come to the opposite conclusion.

The enhanced Tc in single layer FeSe/STO “is caused by the collaboration between an intrinsic mechanism and the interfacial electron-phonon interactions.”

They looked at many samples grown in 3 different configurations: a thick (50 ML) FeSe film dosed with potassium atoms as dopants, grown on STO with the 16O isotope; a single layer (1 ML) FeSe on the 16O STO substrate; and a single layer FeSe grown on STO where the interface is dominated by 18O. The isotope change in oxygen changes the frequency of the interface phonons.

Three FeSe/STO sample configurations
Three sample configurations of FeSe/STO. Figure source.

The ARPES shadow bands were measured as a proxy for the electron-phonon coupling strength at the interface. They quantify this as η, the ratio of the intensity of the shadow band to the main band. They could also extract the superconducting gap values. Their main result is shown below.

Superconducting gap of FeSe/STO and the interfacial phonons
Superconducting gaps of FeSe/STO versus η a proxy for, the strength of electron-phonon coupling with an interfacial phonon mode. Figure source.

The linear dependence between the superconducting gap energy and the electron-phonon coupling agrees with a picture in which electron-phonon scattering is peaked in the forward scattering (q=0) direction, as expected for electron-phonon enhanced superconductivity. The extrapolation to no electron-phonon coupling intersects the gap values of the bulk, potassium doped samples, where the electronic pairing mechanism is understood to be at play. Thus, the authors argue, the enhanced Tc in single layer FeSe/STO “is caused by the collaboration between an intrinsic mechanism and the interfacial electron-phonon interactions.”

So are phonons important or not?

Good question. If I had to place my bets today, I would say that the interfacial phonons absolutely play a role. That is, unless some other shadowy character appears from behind the scenes.

What do you think?

How does FeSe/STO acquire such a high superconducting transition temperature? Any predictions or anecdotes about other systems in which interface physics enhances Tc? Weigh in in the comments!

Featured Image Credit: SLAC National Accelerator Laboratory. Banner Image Credit: arXiv:1710.07057.

Leave a Reply

Your email address will not be published. Required fields are marked *