AmericanChemicalSociety.com

This work focuses on the role of electronic quantum size effects in determining the kinetics and thermodynamics of hydrogen adsorption, dissociation, absorption, diffusion, and recombination in or on metal/metal-hydride nanostructures. Specifically, we are studying these processes on well-defined ultrathin Mg films, which can be made atomically smooth over mesoscopic distances. The experiments are carried out under precisely controlled ultrahigh vacuum conditions. Fundamental understanding of these quantum phenomena is directly relevant to the use of light-metal hydride nanostructures as storage medium for hydrogen.

The kinetic pathways for hydrogen incorporation in ultrathin Mg films indicate a rich parameter landscape in which the film thickness, temperature, hydrogen dose, deposition sequence and the substrate of the Mg layers all play a very important role. Our experimental studies have provided detailed information about hydride formation at the nanoscale. Through our thin film approach, we have been able to acquire the first scanning tunneling microscopy (STM) images of nanoscale hydride formation in ultrathin Mg(0001) films, grown on a Si(111) substrate. Interestingly, our study indicates that hydride growth above and below 10 monolayers (ML) of Mg is strikingly different, as discussed below. In addition, our x-ray photoelectron spectroscopy (XPS) and low energy electron diffraction (LEED) studies indicated the formation of a thin Mg₂Si layer at the Mg/Si interface. The thickness of this interfacial layer decreases with increasing Mg film thickness, indicating that the stability of the buried silicide layer also varies as a function of Mg layer thickness. Mg₂Si is not only important in its own right because of its own potential for hydrogen storage (the presence of silicon reduces the formation enthalpy of the magnesium hydride), but also because the Mg₂Si layer modifies the electronic boundary conditions of the quantum size effect, and thus potentially affects the nucleation and growth of the low-dimensional hydride phase.

Low-temperature (130 K) hydrogen exposure produces a two-dimensional hydride phase that covers the entire surface of the Mg films. Upon further hydrogenation, the planar growth front of the hydride layer moves toward the interior of the Mg film. Evidence for the existence of a two-dimensional ‘hydride skin’ comes from STM and XPS. Whereas the metallic Mg 1s core level signal in XPS increases systematically with the Mg film thickness, the Mg core level intensity associated with the hydride layer is independent of the Mg film thickness. In addition, the surface plasmon loss of metallic Mg is completely absent while the STM images reveal a uniform hydride phase without metallic Mg patches.

Annealing of these hydride films to room temperature, which is well below the desorption temperature measured with temperature programmed desorption (TPD), produces a surprising contrast. First, below 10 ML, the XPS intensity from the hydride layer does not change appreciably after annealing. The STM image shows that the surface is still covered by a disordered hydride skin. However, the two-dimensional hydride layer breaks up above 10 ML, thus uncovering large areas of metallic Mg patches and scattered hydride clusters. Careful fitting analysis of the XPS data indicates a reduced Doniach-Šunjić asymmetry on the main line (compared to pure Mg films), indicating that these metallic Mg patches contain a dilute amount of hydrogen. The composition of the hydride clusters must be close to MgH₂, based on the local band gap values of ~5.5 eV, measured with scanning tunneling spectroscopy. The hydrogen bonding below 10 ML also appears to be stronger, as indicated by our TPD spectra and by the thickness dependence of local hydrogen desorption under a biased STM tip.

Comparison between the post-annealed low temperature deposits and films that have been hydrogenated at room temperature reveals another contrast. Above 10 ML, the resulting hydride morphologies strongly depend on the formation pathway, while below 10 ML, the morphology and stability of the hydride layer are almost independent of the pathway.  For instance, room temperature hydrogenation above 10 ML proceeds via a random nucleation and three-dimensional growth, as opposed to the planar growth observed below 10 ML. We furthermore find that above 10 ML, the Mg films cannot be fully saturated with hydrogen, even after excessive exposure to atomic hydrogen.

In summary, this work produced very detailed nanoscale information about the hydrogenation of ultrathin Mg films and indicated the existence of a ‘critical thickness’ of about 10 ML for the morphological evolution of the hydride phase. Similar critical-thickness phenomena have also been observed in our preliminary theoretical studies of the interaction of hydrogen atoms with freestanding Pd(111) films of varying thickness, showing overall faster atomic H kinetics for thicknesses less than 11 ML. These independent results reveal a common trend in the hydrogen storage and purification aspects of ultrathin metal films. Future studies will focus on the structural and/or electronic origin of the critical thickness phenomenon.