Surprises from NMR in sodium cobaltates

Anonymous

Synopsis

Surprises from NMR in sodium cobaltates

October 20, 2008• Physics 1, s41

After the discovery of superconductivity in doped sodium cobaltate, numerous measurements contributed to mapping out the various magnetic and electronic phases that occur in this material. Now, the report of a new phase diagram may challenge the previous version.

A large part of the interest in superconducting ${Na}_{x} {CoO}_{2}$ is that the ${CoO}_{2}$ layers are reminiscent of the ${CuO}_{2}$ layers in high- $T_{c}$ cuprates, except that the $Co$ ions sit on a triangular lattice, which enhances the role of magnetic frustration between the Co spins. Also like the cuprates, the layered cobaltate ${Na}_{x} {CoO}_{2}$ displays a rich medley of ordered states: superconductivity when intercalated with water for $1 / 4 < x < 1 / 3$ , insulating charge-order at $x = 1 / 2$ , and spin-density wave order at $x = 3 / 4$ . The insulating state at $x = 1 / 2$ separates two distinct metallic states: a paramagnetic metal below $x = 1 / 2$ and a “Curie-Weiss” metal with antiferromagnetically coupled spins above $x = 1 / 2$ .

While no consensus exists on a theoretical picture, this experimental phase diagram is generally believed to be true. Now, Guillaume Lang and colleagues from Laboratoire de Physique des Solides at Université Paris-sud and Laboratoire Léon Brillouin in Saclay report a rather different phase diagram based on nuclear magnetic resonance experiments. Writing in Physical Review B, the authors find that at low temperatures there exists a critical doping range, $x^{*} =$ 0.63–0.65, below and above which antiferromagnetic and ferromagnetic correlations are, respectively, dominant. This contradicts the nonmagnetic behavior reported earlier for $x < 1 / 2$ .

For $0.5 < x < 0.62$ , Lang et al. also identify a doping-dependent temperature scale $T^{*}$ , which separates a high-temperature region with ferromagnetic correlations and a low-temperature region with antiferromagnetic correlations. The $T^{*}$ line slopes away from the insulating limit ( $x = 1 / 2$ ) in the same way the pseudogap crossover line does in the cuprates. The physical origin of this doping- and temperature-dependent crossover in magnetic correlations is a striking new puzzle for theory to address. – Sarma Kancharla