LBNL Masthead A-Z IndexBerkeley Lab mastheadU.S. Department of Energy logoPhone BookJobsSearch
Electron Correlation in Iron-Based Superconductors Print


In 2008, the discovery of iron-based superconductors stimulated a worldwide burst of activity, leading to about two preprints per day ever since. With a maximum superconducting transition temperature (so far) of 55 K, it is natural to wonder if studying the new materials will help uncover one of the deepest mysteries in modern physics—the mechanism of superconductivity in the copper-based "high-temperature superconductors." One clue lies in whether the electrons in the new superconductors are as highly correlated as they are in the high-temperature superconductors. A truly international North American/European/Asian collaboration working at the ALS has now reported results from a combination of x-ray absorption spectroscopy, resonant inelastic x-ray scattering, and systematic theoretical simulations of iron-based superconductors. The team was able to settle the correlations debate by showing that electrons in the iron-based families that were studied favor itinerant (delocalized) states with only moderate correlations.

Are Iron Pnictides New Cuprates?

Owing to their negative electric charge, electrons intrinsically repel each other; however, metals contain large numbers of electrons that do not interact so strongly. Paradoxically, the high mobility and density of electrons give rise to a screening effect that reduces the effective interaction between electrons, so that every electron is both screened and participates in screening other electrons. On the other hand, in some semiconductors and insulators with a low density of charge carriers, the screening effect is much weaker, leading to an electron system with non-negligible interactions. These "strongly correlated" materials often exhibit novel properties, such as superconductivity at much higher temperatures than those of normal metals.

Unfortunately, strongly correlated systems demand a much more complicated theoretical approach. For example, in the two decades since their discovery, a satisfactory theory has yet to be achieved for the famous copper-based "high-temperature superconductors," also known as cuprates. When a new form of iron-based superconductor was discovered in 2008, an active debate followed: would studying the iron-based superconductors shed any light on the mechanism driving high-temperature superconductivity in the cuprates? Working at the ALS, Yang et al. have now provided strong evidence that these new superconductors, although with many similarities to cuprates, are not highly correlated systems. Their results set a clear boundary on theories, and understanding these new materials might lead to technological breakthroughs for optimizing their useful properties.

Layer schematics

Comparing the transition-metal-containing layers in cuprate and iron pnictide superconductors shows that the copper–oxygen layer (top) is flat with relatively large distances between the copper atoms, while the iron–arsenic layer (bottom) is puckered with shorter distances between the iron atoms, giving rise to differences in the electronic structures of these compounds.

The iron-based compounds are called pnictides because they contain a pnictogen; that is, an element from the nitrogen group of the periodic table. Both iron- and copper-based (cuprate) superconductors are layered compounds, and it is believed that the 3d electrons in the respective transition-metal layers play key roles in superconductivity. However, iron, together with cobalt and nickel among the 3d metals, is an archetypal ferromagnetic metal. The magnetic moments are mostly in the 3d bands, and naively, one would expect the well-aligned spin of 3d electrons of iron to prevent superconductivity, which requires pairs of electrons with opposite spin directions. Another crucial difference is that all the iron 3d orbitals contribute charge carriers, while for cuprates a single half-filled band dominates the important physics.

While a half-filled band signifies a metal in conventional band theory, the electrons in cuprates are also strongly correlated, owing to the strong Coulomb interaction, which prevents two electrons from occupying the same site, resulting in a so-called Mott insulator. The lack of information on the strength of electron correlation in the iron pnictides has blocked the way toward a consensus for the minimal model needed to describe the electron pairing mechanism in these materials. Theoretical results have been controversial.

To explore this issue, the collaboration performed x-ray absorption spectroscopy (XAS) and resonant inelastic x-ray scattering (RIXS) measurements at ALS Beamline 8.0.1, where they investigated five iron-containing materials, including an iron pnictide superconductor (SmO0.85FeAs) with a record high 55-K transition temperature, two non-superconducting iron pnictides (BaFe2As2 and LaFe2P2), and for comparison iron metal and an Fe2O3 insulator. The spectra of iron pnictides exhibited qualitative, in some cases quantitative, similarities to those of the iron metal but showed no features resembling the multiple peak structures seen in iron-based insulators. Furthermore, a RIXS study across the resonant XAS edges demonstrated that the resonance spectra are dominated by the nonresonant "normal fluorescence," with no observance of excitation peaks. The team interpreted these results as showing the importance of the iron metallicity and strong covalency in these new iron-based superconductors.


Left: XAS (above) and RIXS (below) spectra of an iron pnictide superconductor with a transition temperature of 55 K (the same type of spectra were obtained for the other four samples studied). RIXS spectra were taken at excitation energies marked by the arrows, which range across the iron absorption edges, as indicated by the numbers in the XAS spectra. Right: Comparison of RIXS (left) for an excitation energy of 708 eV and XAS (right) spectra for the five iron-containing materials studied. The RIXS energy loss is the difference between the excitation and emission energies. The Fe2O3 spectra are distinctively different from those of the iron metal and the iron pnictides.

The team then turned to a systematic theoretical study to simulate the experimental results, and more important, to pin down the upper limit of electron correlation in the new superconductors. They first performed calculations based on a Hubbard model (the simplest model in solid-state physics of interacting particles on a lattice). The calculations suggested a relatively minor role for correlations in the iron pnictides. Subsequently, the comparison between different theoretical models and experimental data indicated that, instead of localized states due to strong electron interactions, electrons in iron pnictides prefer itinerant states with moderate correlation strength.

Calculated  spectra

Calculation of the x-ray absorption spectra for iron pnictides based on a Hubbard model with various values of the on-site Coulomb repulsion U. An upper limit of U of about 2 eVwas determined by comparing these spectra to the experimental data. This value of U is comparable to or even smaller than the electronic band width, indicating these iron-based superconductors are not strongly correlated systems.

These results will help lead physicists to the mechanism of superconductivity in iron pnictides and perhaps also to their optimization for technological applications.



Research conducted by W.L. Yang, J. Denlinger, and Z. Hussain (ALS); A.P. Sorini, B. Moritz, and W.-S. Lee (SLAC National Accelerator Laboratory); C.-C. Chen, J.-H. Chu, J.G. Analytis, I.R. Fisher, Z.-X. Shen, and T.P. Devereaux (SLAC National Accelerator Laboratory and Stanford University); F. Vernay and B. Delley (Paul Sherrer Institut, Switzerland); P. Olalde-Velasco (ALS and Instituto de Ciencias Nucleares, Mexico); Z.A. Ren, J. Yang, W. Lu, and Z.X. Zhao (National Laboratory for Superconductivity, China); and J. van den Brink (SLAC National Accelerator Laboratory and Leiden University, The Netherlands).

Research funding: U.S. Department of Energy (DOE), Office of Basic Energy Sciences (BES); Stichting voor Fundamenteel Onderzoek der Materie (FOM), the Netherlands; and Consejo Nacional de Ciencia y Tecnología (CONACyT), Mexico. Operation of the ALS is supported by BES.

Publication about this research: W.L. Yang, A.P. Sorini, C.-C. Chen, B. Moritz, W.-S. Lee, F. Vernay, P. Olalde-Velasco, J.D. Denlinger, B. Delley, J.-H. Chu, J.G. Analytis, I.R. Fisher, Z.A. Ren, J. Yang, W. Lu, Z.X. Zhao, J. van den Brink, Z. Hussain, Z.-X. Shen, and T.P. Devereaux, "Evidence for weak electronic correlations in iron pnictides," Phys. Rev. B 80, 014508 (2009).


ALSNews Vol. 306