Physics of valence fluctuation as exemplified by Eu-based materials

October 24, 2018

MPI-CPfS co-workers: S. Seiro, C. Geibel, J. Sichelschmidt, V. Guritanu, N. Caroca-Canales

Strongly correlated electron systems are one of the fascinating topics in modern solid state physics, and certainly one of the most intriguing. In these systems the valence electrons, which are responsible for most of the properties of a solid, interact very strongly with each other in such a way that they cannot be thought of be as individual particles, as e. g. in classical semi-conductors like silicon. Instead they form collective states, with properties very different from those of the constituent electrons and which frequently defy standard concepts of solid state physics. The origin of such peculiar behavior is in most cases the interaction between “localized” electrons, which are bound to a specific atom, and “itinerant” electrons, which are free to move within the solid, or to electrons at the border between “localized” and “itinerant” behavior. The complexity of the collective electronic states formed in such systems makes both their theoretical description and their experimental determination very challenging.

The aim of the present project is to obtain a deeper insight into a category of strongly correlated electronic systems which is far from being fully understood: europium-based valence fluctuating compounds. The element europium (Eu) is one of the rare earth metals which are of strong relevance not only for fundamental research, but also for technological applications. But Eu is quite special compared to other rare earth elements because it can easily switch from a fully trivalent, non-magnetic state to a fully divalent, magnetic state, depending on chemical composition of the compounds or external parameters like temperature (T), magnetic field or pressure. Therefore Eu-based intermetallic compounds are prototypical examples for the so-called valence fluctuating systems where strong electronic correlations emerge from large charge fluctuations. The essential term in their electronic Hamiltonian is the Coulomb repulsion Ufd between a localized f-electron and a valence d-electron on the same Eu-atom. Neither the reason why Ufd is so large for Eu systems compared to other ones nor the way it influences the physical properties is well understood.

One of the reason for the absence of in-depth studies of Eu based valence fluctuating systems is the difficulty of growing appropriate high quality single crystals. In order to reach the valence fluctuating regime, one needs to combine Eu, which has a comparatively low boiling point of 1596 °C, with high melting transition metals like e.g. Rhodium (Rh) (Tm = 1964 °C) or Iridium (Ir) (Tm = 2446 °C).  The resulting high vapor pressure of Eu, its very high reactivity, and the high melting temperatures Tm of the compounds of interest makes the growth of high quality single crystal challenging. Profiting from our comprehensive experience on the crystal growth of the Yb based homologue compounds YbRh2Si2 and YbIr2Si2, we adapted and optimized our flux growth technique to the more reactive Eu and were able to grow sufficiently large and high quality single crystals of the compounds EuCo2Si2, EuIr2Si2 and EuRh2Si2. This series, with isovalent transition metals, covers the whole range from trivalent Eu (EuCo2Si2) to valence fluctuating Eu (EuIr2Si2) and divalent Eu (EuRh2Si2). In addition we also obtained high quality single crystal of the valence fluctuating Eu system EuNi2P2.

In a first step we determined the optical conductivity of the two valence fluctuating systems EuNi2P2 and EuIr2Si2. The optical conductivity s(w) is a particularly interesting property because it tracks the effect of correlations on charge excitations within a very large energy range, usually from a few meV to a few eV. Our spectra, shown in Fig. 1, represent the first s(w) results ever obtained on valence fluctuating Eu-systems. The correlations manifest themselves in a strong temperature dependence of s(w), and in the appearance of a pronounced structure in its frequency dependence at low T. Above 200 K s(w) shows only a weak and continuous increase with decreasing w. But below 100 K a clear maximum and a clear minimum form at around 150 meV and 20 meV, respectively. Furthermore, including results of dc (w = 0) conductivity measurements shows that there is a strong increase of s(w) for w < 20 meV at low temperatures. This increase reflects the formation of conducting collective electronic states with a mass of the order of 10 times the free electron mass. On the other hand the nature of the peak at 200 meV is not clear. A simple model for valence fluctuating systems, the Falicov-Kimball model, predicts such a maximum at an energy corresponding to the on-site Coulomb repulsion Ufd, but a similar maximum is also observed in Ce or Yb-based intermediate valent systems where it corresponds to the hybridization energy between localized and valence electrons. Likely it origins from both interactions, but there is presently no appropriate theoretical prediction. Our results have stimulated new theoretical efforts on models including both hybridization and Ufd and we hope to obtain further insight in the near future. In a next step we have meanwhile determined the Fermi surface of EuIr2Si2 using de Haas van Alphen experiments and are analyzing the data. This would be the first determination of the Fermi surface of a valence fluctuating Eu system.

In the course of studying these Eu compounds we discovered an interesting feature which is relevant for a completely different topic, namely surface science and spintronics. We performed extended Angle Resolved Photoemission Spectroscopy (ARPES) measurements on EuRh2Si2. This compound has a magnetic divalent Eu state which orders antiferromagnetically below 25 K. We revealed that the Si-Rh-Si surface trilayer of a Si terminated surface hosts an electronic surface state with an unexpected and large spin splitting. This splitting is induced by the underlying magnetic Eu layer, but surprisingly, it sets in at a temperature of 32.5 K, well above the bulk ordering temperature of 25 K. In reciprocal space this surface state is located within a gap of the bulk electronic states and thus constitutes a two-dimensional electron gas that may be polarized through an underlying magnetic layer. We are now studying this feature in homologous compounds with other rare earth elements which will allow us to increase the ordering temperature or to induce a strong magnetic anisotropy.

In cooperation with:  S. Kimura1, D.V. Vyalikh2, and C. Laubschat2

1UVSOR Facility, Institute for molecular Science, Okazaki 444-8585, Japan

2Institute of Solid State Physics, Dresden University of Technology, D-01062 Dresden, Germany    

 

Further reading:

[1] Optical Study of Archetypical Valence-Fluctuating Eu Systems; 
     V. Guritanu, S. Seiro, J. Sichelschmidt, N. Caroca-Canales, T. Iizuka,
     S. Kimura, C. Geibel, and F. Steglich, Phys. Rev. Lett. 109, 247207 (2012).
     MPG.PuRe 

[2] Strong ferromagnetism at the surface of an antiferromagnet caused by
     buried magnetic moments
; A. Chikina, M. Höppner, S. Seiro, K. Kummer,
     S. Danzenbächer, S. Patil, A. Generalov, M. Güttler, Yu. Kucherenko,
     E. V. Chulkov, Yu. M. Koroteev, K. Koepernik, C. Geibel, M. Shi, M. Radovic,
     C. Laubschat, D. V. Vyalikh, Nature Communications 5, 3171 (2014).
     MPG.PuRe

Go to Editor View