Reactivity at the nanoscale


Catalysis takes place at surfaces. This single detail has pushed Surface Science research into studying the interaction of molecules with well defined surface. Furthermore, the reactivity of materials can change radically when we talk about a few atoms. For example, gold can be a good catalyzer in nanoparticle form. In our group we have been studying the effect of thickness in very thin metallic films on the reactivity towards simple molecules.  In particular, we have studied the interaction of hydrogen with magnesium, cobalt and palladium films. This work, mostly the PhD thesis of B. Santos, show the range of fates that the hydrogen can end up in a metal film (his PhD thesis is also where most of the following stuff comes from).

Magnetism at the atomic layer limit


We work on magnetic films, and their properties when their thickness is of the order of a few atomic layers. When you go down to such level of detail, you can find new phenomena. Is a single layer ferromagnetic? How does its properties change with thickness? What about a single layer of an antiferromagnetic material? This problems have been the subject of research in the last 50 years. Surprisingly, there is still a lot we can learn with the right combination of techniques.

As an example, and given the recent Nobel prize on physics for the discovery of the giant magnetiresistance that we use in our hard-drives. The original material combination where the effect was discovered was chromium and iron. Surprisingly enough, we still did not know whether a single layer of chromium by itself is magnetic or not. We recently used a series of techniques to find out that even a single layer actually presents antiferromagnetic order.

As another example, consider the easy-axis of magnetization of a ferromagnetic material: the direction along which the magnetization is oriented in a sample in the absence of an applied magnetic field. The magnetization direction of the film can depend on the film thickness. We recently found that atomic films just one atomic layer thick of cobalt on a substrate of ruthenium has the easy-axis of magnetization in the plane of the film. That is actually expected for a thin film. But surprisingly enough, a film -or islands- of two atomic layers changes the magnetization (i.e., has a spin-reorientation transition) to an out-of-plane orientation. coru spleemAnd films three atomic layers (and thicker) change again their magnetization direction (the figure shows 3 atomic layer islands on top of a two atomic layer film. Of course, the only way to see this kind of behaviur is look close enough. Cobalt on ruthenium, such many other materias, is very difficult to grown in perfect layers, so we can observe this effect. In fact, only by looking in terraces of the substrate several micrometers wide we could find islands perfect enough to do our experiment.

The technique we use to locally image the magnetic domains in a given direction is spin-polarized low energy electron microscopy. This is a low energy electron microscope that employs a beam of electrons that is spin-polarized. Reflection of the beam from a ferromagnetic sample produces some contrast due to exchange scattering between the sample and the beam electrons. By subtracting images with opposite spin polarizations, we can map the magnetic domains of the sample. There are only a few systems in the world. We did this work in collaboration with Andreas Schmid at Berkeley lab.

And schematic overlaid on a topograhic image can be show here. This work was selected as a Physics Review Focus story, you might want to look at it there.

We have also tried to cover the Co films with different materials, such as Cu, Ag, and Au. And we have found that also the growth of a non-magnetic capping layer produces spin-reorientation transitions. As an example, covering the cobalt films with silver we found the following changes in the magnetic easy axis, the arrows indicate the magnetization direction of films of Co with a given coverage of each of the coin-metals. As you can see, we have found quite a few combinations that present out-of-plane magnetization.




Self-organization in metal growth


By means of playing around with competing interactions, it is possible to generate indirectly (or by a bottom-up approach) ordered patterns on the nanoscale, patterns that could well be used to grow additional material. But to do so, we must learn about those interactions. Depending on your point of view (or your hype factor) this is just the usual way of doing things, or the next big thing in nanotech. To be fair, self-assembly and self-organization is around us in everyday life. Blacksmiths have hundreds of years of experience modifying the properties of iron and steel to obtain the desired microstructure for a given function. It is only in the late XX century that we have got used to the top-down approach, lithographic techniques in microelectronics that allow us to put every gate, every transistor in the desired location. But this mode of creating a new material (or device) is still the exception.

We focus on surfaces. You can find more information in the Master's short course that you can find here. I would like to point out that the future in many fields, and magnetic recording media is a clear case, lies in the combination of both lithographic techniques with self-assembly. Nowadays Seagate and Hitachi are exploring the use of the self-assembly of diblock copolymer combining the good short range order of self-assembly with the long range control in lithography.

From a basic point of view, a particularly simple -and common- way to promote self-assembly on surfaces are competing interactions of different range. Say for example a short range attractive interaction (due to broken bonds), and some sort of long range interaction. The origin of the long range interaction can be quite different: magnetic dipolar in thin films with perpendicular magnetic anisotropy, electrostatic interaction between polar molecules, or elastic interactions due to the stress difference between surfaces regions (be it by adsorbates or by structure). This produces patterns of the type shown in the figure. A good example of a pattern produced by the latter effect is the one that arises when depositing gold on tungsten at high temperature, where we found that the pattern formation can be understood in a particularly simple model.


Also the generation of misfit dislocations (which can be considered a competition between intralayer interactions vs interlayer interactions) has often been used to produce well ordered patterns, like the nice dislocation pattern we observed (and explained) on Ag(111)/Ru(0001) More information here.

A different way to tackle self-organization is to take advantage not of the thermodynamics of the system (i.e. the equilibrium structure) but kinetic limitations. Maybe tuning the temperature we can take advantage of some activated process on surfaces to self-limit island growth. Such an example is the is the serpentine growth of Pd on Ru, were the growth of Pd self-limits due to some alloying with the substrate at steps edges that limits further growth.