LSU Surface Science Research

Half-Metallic Oxide Thin-Films

We are interested in thin-films which are useful for computer disk head sensors and MRAM. Half-metallic materials such as Fe₃O₄ have one spin type at the Fermi level and hold great promise in devices based on spin-dependent transport. In principle, a spin-dependent conductor can provide enhanced sensitivity to magnetic field. We have focused on thin-films of this oxide grown on Cu substrates, for obvious device-related reasons.

Cleaning the Cu(001) Substrate

The Cu(001) crystal is cleaned by cycles of Ar+ sputtering and annealing. The image on the left shows a sputtered surface: clean but the landscape is closer to that of the moon than that of a flat surface. Annealing produces large flat terraces with monoatomic step heights of ~ 1.8 A.

Growing Fe on Cu(001) There are many observations in the literature regarding the growth of fcc-Fe on Cu. The kind of film one gets depends on many parameters which vary with deposition technique (evaporated, pulsed-laser deposited, etc.) The group that has spent most time investigating this is Kirschner's group (Julich). What we show here is sub-monolayer Fe on Cu. In the left panel the substrate was ~50 C and dark pits form which are probably due to an intermixing of Fe with Cu at the surface. If the substrate is at room T, the growth is essentially layer-by-layer but many authors have noted that this really depends on vacuum conditions.

Oxidizing <2 ML Fe on Cu(001) We identify the diagonal strip of oxide in the left panel as FeO(111). It has remarkably good epitaxy, grows it 5A thick layers and flattens the Cu surface locally. There are huge flat Cu terraces with monatomic Cu steps and the Cu remains atomically clean. In photoemission, we can see that the surface states of Cu(001) are very strong. The oxide has dark stripes separated by 360 A which we identify as antiphase domain boundaries where strain is relieved. A closer look in the right panel also shows a hexagonal superstructure (~0.2 A corrugation) which we can understand based on our structural model for the oxide.

Oxidizing thicker Fe on Cu(001) For more than 2 ML of Fe, we form large crystallites of Fe₃O₄(111) that are oriented 15 degrees from the Cu [010]. This also has dark stripes indicating strain relief but they are separated by 170 A. In the right panel, we can also see a hexagonal superstructure but the orientation of this is rotated by 30 degrees relative to the stripe direction that we saw for FeO(111). Photoemission studies show that the surface is not completely covered with oxide, even for very thick initial Fe films. We are currently investigating this. Complete surface coverage can be achieved by several cycles of Fe deposition and oxidation.