The researchers, led by Shahal Ilani, made their new scanning charge detector by placing a nanotube at the end of a scanning probe cantilever. The nanotube acts as a quantum dot, or single electron transistor (SET), in which the number of electrons on the nanotube, and thus the electrical current flowing through it, can be precisely controlled using nearby gate electrodes. "Indeed, the current through the nanotube is very sensitive to electrostatic charges in its surroundings, so by measuring this current as a function of a detector's position above the LAO/STO interface, for example, we can image the local electrostatic landscape,” explained team member Joseph Sulpizio. “We can also measure changes in the local mechanical properties of a material (for example, its topography and its piezoelectric response) by applying a potential difference between the material and the SET, and measuring changes in the local capacitance between the two.”

The new technique is particularly good at imaging “buried” interfaces like those between STO and LAO – “hidden” from view to standard surface probes like scanning tunnelling microscopy that rely on electron tunnelling and which are insensitive to electronic signals coming from deeper within a sample.

“Our scanning probe can measure both the electrical and mechanical properties of a material at the same time, which means that we can measure the topography of the LAO/STO interface, its local piezoresponse and its electric potential all at once,” Sulpizio told “The images created from these separate channels consistently show that there are different crystal domains in the underlying STO that we can manipulate with gate voltages and which have a clear electrostatic fingerprint.”

Strontium titanate (STO) is a wide-gap insulator with a perovskite structure that is routinely used as a substrate for growing high-temperature-superconducting cuprates, colossal magnetoresistive manganites and multiferroics, among other materials. In 2004, researchers discovered that the interface between a thin layer of LAO and STO contains a quasi-two-dimensional conducting electronic system. Later on, they also found that the interface had a number of unique properties, such as gate-tunable magnetism, superconductivity and spin-orbit interactions.

Puzzlingly giant piezoelectricity

“Our experiments have now settled the debate concerning the origin of the puzzlingly giant piezoelectricity in STO,” said Sulpizio. At low temperatures, the cubic crystal of STO undergoes a structural (ferroelastic) phase transition to a rectangular prism crystal structure, he explains. These prisms have one long axis and two short axes, and the unit cells in a STO sample can thus either be oriented with the long axis in the plane or perpendicular to it.

Different domains form within the crystal and they can be characterized by their orientation – either in- or out-of-plane. “We found that the different domains inside STO couple to applied gate voltages such that the orientation of the unit cells at a given position in the crystal can be changed as the gate voltage varies,” says Sulpizio. “Thanks to the different unit cell side lengths, this change in orientation causes a change in the STO surface height, demonstrating that the large piezoelectricity observed in this material is in fact related to the motion of microscopic domains”

Striped pattern creates a striped electric potential landscape

And that is not all: the Weizmann researchers also showed that the structural domains in STO form a striped pattern that creates a striped electric potential landscape at the LAO/STO interface. “We reckon that this striped potential should influence the density of electrons at the interface, which in turn should strongly affect how current flows through devices made from LAO/STO,” explained Sulpizio. This striped current flow was, in fact, recently and independently measured by a group of researchers at Stanford University in the US.

The Israel team says that past charge transport experiments in LAO/STO could now be re-examined in the light of these new results and future experiments should be designed with the striped domains in mind. “There may be very interesting physics at play at the domain walls themselves, and, who knows, we may even be able to engineer domain patterns in LAO/STO to create novel nanodevices, such as 1D channels,” added Sulpizio.

The researchers say that they are now busy looking at how to electrically manipulate the STO domains, and image their electric potential with higher resolution to distinguish between the physics taking place at the bulk of the domains and at the domain walls.

These experiments were performed in collaboration with Eli Zeldov's group, also at the Weizmann Institute, and the work is reported in Nature Materials doi:10.1038/nmat3810.

Further reading

Phenomenological model uses CAFM to characterize tunnel-junction-like devices (Nov 2012)
Unexpected heterostructure seen in STO (Aug 2011)
Electron gas conducts at insulating interface (Feb 2013)