Magnetism and superconductivity in layered cobaltates

In 2003, superconductivity was discovered in Na_xCoO₂·yH₂O, a material with a structural backbone of triangular CoO₂ layers. Next to the cuprates and Sr₂RuO₄, this is the third transition metal oxide with quasi-two-dimensional electronic structure that exhibits this phenomenon. Motivated by our work on the cuprates, as well as by theoretical work in our group that predicts unconventional superconductivity in the cobaltates [1], we embarked on a project to prepare high-quality single crystal samples in collaboration with the crystal growth group [2]. The resulting specimens have allowed us to characterize the spin and charge excitations of this material, with a focus on metallic but non-superconducting samples with y = 0 and x > 0.5.

Soon after the discovery of superconductivity, evidence began to emerge of a magnetically ordered state in non-hydrated samples with large Na content. Our group was among the first to prepare homogeneous samples with this composition. Thermodynamic and muon spin rotation experiments established a magnetically ordered state below T_N ≈ 20 K encompassing the entire sample [3]. Spin-polarized neutron diffraction was used to determine the magnetic ordering pattern as that of an A-type antiferromagnet, that is, the exchange interactions within the CoO₂ layers are predominantly ferromagnetic, and the ferromagnetic sheets are antiferromagnetically coupled [4]. Surprisingly, however, the spin wave spectrum, also determined by neutron scattering, revealed that the magnitude of the antiferromagnetic interlayer coupling is comparable to that of the ferromagnetic intralayer coupling.

Spectral features in infrared ellipsometry experiments on the same samples provided striking evidence of polaronic excitations [5] (Fig. 1a). The spectral weight of the polaron features is sensitive to the onset of magnetic order. These data could be interpreted in terms of a microscopic model in which mobile holes modify the spin state of surrounding Co ions, thus generating an extended magneto-polaron mode. A subsequent refined theory devised in collaboration with scientists in Metzner's group explains the thermodynamic data as well as the experimentally observed magnetic order and dynamics as consequences of magneto-polaron formation [6] (Fig. 1b).

Because of the large lattice expansion associated with water intercalation, the preparation of high quality hydrated specimens has proven difficult. The first results have come from electronic Raman scattering, where polaron features akin to those observed in the infrared experiments can be seen for unhydrated samples with x ≈ 0.8 [7]. In the superconducting range of the phase diagram, these features disappear, and a large pseudogap opens up at temperatures far exceeding the superconducting transition temperature. This observation is particularly significant in view of the pseudogap in the cuprates, the origin of which has long been debated. Phonon anomalies with an onset temperature comparable to that of the pseudogap suggest that in the cobaltates this phenomenon is associated with charge order. The observation of charge order in superconducting samples is important for theories of the origin of superconductivity in this class of materials. In collaboration with the technology service group, we have recently succeeded in preparing high-quality epitaxial thin films of layered cobaltates [8]. This will have considerable implications for future work in this field. Thin film samples can be prepared over a wide range of doping levels and exhibit superior surface quality, as demonstrated by initial infrared ellipsometry experiments. Moreover, superconductivity can be induced in these specimens without significant degradation of the sample quality. In order to fully explore the potential of these films, we have established a collaboration with Profs. Lambert Alff (Technical University of Darmstadt) and Peter Lemmens (who recently moved from the MPI-FKF to the University of Braunschweig).