Research

Edge magnetism in CrGeTe3: when the tail shakes the dog

CrGeTe3

Magnetism is a phenomenon that emerges from the interaction between the spin degree of freedom of neighboring electrons. This collective phenomenon typically weakens in reduced dimensions. For example, near the surface, electrons have fewer neighbors; thus, the constraint on the electron spin is lower. Therefore, surface electrons are usually paramagnetic or have lower coercivity than bulk electrons. We recently observed the opposite situation in CrGeT3 thin films. For samples thicker than 10 nm, the region of the sample far from the edges displays random magnetic domains with zero net magnetization at zero field. In contrast, the edges remain fully magnetized in the direction corresponding to the last field applied to the sample. This phenomenon was discovered by direct magnetic imaging (see images below and related article).

CGT edge data1

Strikingly, in quasi-one-dimensional samples, the edge magnetizes the interior. Here the negligible part of the sample, the edge, influences the properties of a more significant part of the sample, the interior. It is the equivalent of observing a tail shaking a dog. This phenomenon is illustrated in the image below. The figure below depicts tge magnetic image of the stripes. The images of the wider stripes include two distinct magnetized edges (color-coded in red) separated by a zero-average magnetization in the stripe’s interior (color-coded in green). However, below a certain critical width, these two edges appear to merge to form a single magnetic domain. Panel c represents a line profile of the image shown in panel b along the x-axis. We note that the magnetic signal for the narrow stripe is four times larger than the signal at a single edge. This finding suggests that the sample interior also becomes a hard ferromagnet because of the edge proximity (link to article).

CGT data 2

 

Magnetism down to the single layer limit

CrSBr

In some materials, each layer is ordered ferromagnetically, but each layer's spin points in the opposite direction compared to its neighboring layers (see blue and red arrows in panel a of the figure below). These materials are called anti-ferromagnetic. Given the collective nature of this phenomenon, we ask ourselves, what happens when reducing the dimension of the sample? Using our microscope, we investigate the material's magnetic properties as a function of the number of atomic layers down to the single-atomic layer. We showed that this anti-ferromagnetic material shows no magnetic field when it has an even number of layers because each layer cancels the other perfectly. For an odd number of layers, we measured the magnetic field of the uncompensated atomic monolayer (panel b). We also show that magnetic domains are present in this single-atomic layer (panel c). (link to article)

CrSBr data

(a) Illustration of our measurement setup. The SQUID-on-tip (SOT) probes the magnetic field component in the c direction. (b) Image of the c component of the magnetic field for a sample with 3 layer. (c) Magnetic domain formed in the same sample showed in b.

 

Superconducting Vortex Manipulation

Vortex manipulation

 

Abrikosov vortices have long been considered as means to encode classical information in low- temperature logic circuits and memory devices. Although it is possible to control individual vortices using local probes, scalability remains challenging. Vortex logic devices require means to shuttle selected vortices reliably over long distances between engineered pinning potentials. At the same time, all other vortices should remain fixed to their precise locations. Here we demonstrate such capabilities using Nb loops patterned below an NbSe_2 layer. SQUID-on-Tip (SOT) microscopy reveals that the loops can position vortices in sites designated to a precision better than 100 nm; they can realize “push” and “pull”  as far as 3 \mathrm{\mu}m. Successive application of such operations shuttles a vortex between adjacent loops. Our results may be used as means to integrate vortices in future quantum circuitry. Strikingly, we are able to demonstrate a winding operation. Such winding, if realized in topological superconductors, is considered an essential part of future topological quantum information processing. (link to article see also video)