Nr 14 / cover - MaNEP news - Scientific Highlight - Publications - Publications 2 - Calendar
Scientific Highlight - page 3
Superconductivity at the interface between insulating oxides by Nicolas Reyren, UniGE

Interfaces between complex oxides may display unusual and novel electronic properties. The interest in the physics of such interfaces has been largely boosted by the 2004 discovery of Ohtomo and Hwang of a metallic interface between LaAlO3 and TiO2 terminated SrTiO3 1. It was also found that the mobility at this interface can be very large. A possible explanation for this phenomenon is the so-called “polar catastrophe” scenario which may occur when the LaAlO3 thickness exceeds a critical value 2 .

This phenomenon is related to the polar nature of the LaAlO3 atomic planes. When these are stacked on top of each other an electric potential develops as the film thickness increases. Above a critical thickness, an electronic reconstruction may take place which could lead to a charge transfer at the interface (thus explaining the observed conductivity), creating a quasi two dimensional electron gas (see Figure 1).

We recently discovered that the ground state of this electron gas is a superconducting condensate. Remarkably, the interface between these two insulating oxides is superconducting! (See Figure 2.) The superconducting transition temperature is about 200 millikelvin and the observed characteristic superconducting properties reflect the two dimensional nature of this system. More specifically, we looked for the footprint of a Berezinskii-Kosterlitz-Thouless (BKT) transition expected in a 2D superconductor.

In the BKT framework, one expects that the voltage versus current curves follow a power law behaviour V∞ /a

 with the exponent a jumping from 1 to 3 at the BKT transition temperature. In a real system with a finite size, the transition is smoothed. However, the transition temperature is, as in the ideal case, defined by the relation a(TBKT)=3. Independently, the resistance versus temperature curve should follow a particular behavior:

R=R0∙exp(−b(T/TBKT−1)−1/2).

Our transport measurements are consistent with these expressions. Both of these predictions have been checked for an 8 unit cell sample (grown in the group of J. Mannhart, in Augsburg) and give us remarkably similar TBKT values, close to 200mK (see Figure 3).

The mechanism behind superconductivity is not yet determined. Two models (not exclusives) are invoked: the electrons at the interface may pair because of the highly polarizable SrTiO3 background, or oxygen vacancies may be present close to the interface, doping SrTiO3 (which is superconducting). In this latter scenario, recent electrostatic field effect measurements suggest that the superconducting sheet is amazingly thin, only a few unit cell thick.

Figure 1: The “polar catastrophe” may lead to an electronic reconstruction.

Figure 2: Superconducting transitions of a series of samples grown in Augsburg and in Geneva.

[ click figure to enlarge ]

Figure 3: V-I characteristics, a(T) and R(T) are consistent with a BKT transition.

[ click figure to enlarge ]

1 A. Ohtomo, H. Y. Hwang, Nature 427, 423 (2004).
2 S. Thiel et al, Science 313, 1942 (2006).
3 N. Reyren et al, Science 317, 1196 (2007).
Nicolas ReyrenNicolas Reyren works in the group of Prof. Jean-Marc Triscone at
UniGE, Department of Condensed Matter Physics (DPMC)