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).

In the reaction of hydrogen and a metal, several steps occur before a metal hydride is formed. First, the hydrogen molecule is adsorbed on the metal surface. The hydrogen molecule has a large dissociation energy of 4.52~eV/molecule. If the hydrogen atoms that form the molecule are going to travel into the metal, the molecule has to dissociate into two hydrogen atoms. This sometimes involves crossing an energy barrier (in which case the dissociation is said to be activated), and sometimes can happen without a kinetic barrier (non-activated dissociation). The most energetically favorable position for hydrogen on a metal is often the adsorbed position on the surface. In some metals, there are sites below the surface that have a similar energy to the adsorbed positions (although in most metals those positions have a higher energy). For example, for Ni, the adsorbed hydrogen atom must overcome a 24 kcal/mol (1.03 eV/atom) potential barrier to diffuse inside the metal, much larger than the average energy of hydrogen at RT. However, if hydrogen has enough energy to overcome that potential barrier, it can diffuse into the metal. Once the atomic hydrogen is inside the metal, the equilibrium lattice position is determined by the potential energy. Most metals can incorporate a given concentration of hydrogen as a solid solution, due to the small size of the hydrogen atom. This is usually called the alpha phase of the corresponding metal hydride. It produces a slight expansion of the lattice of the host metal. But much larger hydrogen concentrations are possible, as often ordered compounds of the metal and hydrogen exist, such as PdH or MgH2. The phase transformation from the diluted alpha phase to an ordered hydride is a first order transformation in which the lattice spacing of of the hydride can change abruptly by a large fraction. In a bulk material, this can fracture it.
The first problem encountered when planning research on hydrogen interacting with metal surfaces is how to detect it. As hydrogen has one electron (in neutral form), its scattering power is much smaller than metal atoms so a direct observation by low-energy electron diffraction (LEED) is often impossible. Typical electron spectroscopies do not fare much better, as Auger electron spectroscopy or X-ray photoelectron spectroscopy are also blind to hydrogen. Even in scanning tunneling microscopy, hydrogen detection is very difficult. The problem is complicated by the fact that most surface science experiments take place in an ultra-high vacuum that is composed mostly of hydrogen. A reactive surface at a pressure of 10^-10 Torr, which is usually considered a ``good'' vacuum, will be covered with hydrogen in less than three hours.

However, hydrogen absorption on metals produces structural or electronic changes in the metal atoms that might be simpler to detect than hydrogen itself. For example, as mentioned before, hydrogen absorption induces an expansion of the lattice parameter of the crystal, which is typically proportional to the amount of hydrogen incorporated in the lattice. This suggests that crystallographic techniques capable of measuring the average lattice spacing of a film, such as low-energy electron diffraction or surface X-ray diffraction can be used to follow the fate of hydrogen in a metal film.

Hydrogen also produces changes in the electronic structure and the magnetic properties of metals. Hydrogen does not have core levels, but it changes the valence band of the metal. Greuter et al. showed that the metal d band shifted downwards in energy respect to the Fermi level because of increased bonding with the hydrogen 1s band. Finally, both the changes in the electronic structure of a ferromagnetic metal or its strain can change drastically its magnetic properties, opening another method to follow the incorporation of hydrogen in a metal film. For example, hydrogen induces an spin reorientation transition on ultrathin Ni films. Other work reported that the magnetic moment of magnetic surfaces is reduced upon hydrogen adsorption.

So far we have studied the response to hydrogen exposure by three different metallic films, Pd, Co and Mg on the same substrate, Ru(0001). The substrate is chosen because hydrogen does not go into the bulk of Ru nor it easily forms an hydride. Furthermore, it is easy to clean by flashing and annealing cycles, and it is a substrate that does not alloy strongly with the metals to be grown on top, at least at the temperature range where the metals show islands large enough to be detected in low-energy electron microscopy. The three metal films selected span a range of behavior with hydrogen: Pd is the prototypical metal that absorbs hydrogen, Mg forms an stable ordered hydride and Co films supposedly only adsorb hydrogen on the surface. All the experiments described were performed under ultra-high vacuum conditions. The metal films were grown by molecular beam epitaxy in the same chamber where their structure was characterized, and they were later exposed to hydrogen. Hydrogen exposure was performed by filling the experimental chambers with molecular hydrogen, or in some cases, by using an atomic hydrogen source.