The NANOAM project is a collaboration between US scientific research groups (funded by NSF) and European scientific research groups (funded by the EC). The goal is the development of scientific understanding of interface-stabilized, equilibrium, thin films that have been observed in some materials, especially ceramic materials. While these thin films have been observed, their behaviors have not been fully characterized and their scientific explanation has not been satisfactorily realized.

Because these films are very thin and because the thickness of these films can be engineered with chemistry and because these films have novel physical properties, these scientific investigations may result in the use of these films as enabling nanomaterials in new devices such as thin film dielectrics.

A very simplistic illustration of the concept follows:

Illustration of an thin film that is a few tens of nanometers thick in equilibrium with a crystalline ceramic. The film has a structure and a composition that differ from the surrounding bulk glassy material (in blue)---and yet it is in equilibrium with the surrounding material and the crystalline, ceramic grains (illustrated as rectangular blocks). Because its structure and chemistry differ from the bulk's, the thin film's physical properties, such as optical and electrical, also differ from the bulk. Furthermore, the thickness of the film depends on both the chemistry of the bulk (blue) material and the crystalline ceramic material. The ceramic crystals above and below the thin film are in different crystal orientations. Such films have been observed, for example, at silicon nitride and strontium titanate grain boundaries.

Similar, stabilized, thin films have been observed on the exterior surfaces of ceramic materials. One goal of this project is to understand such films so that they can be engineered for devices.

These films present a large number of behaviors that need to be characterized and explained. Such questions include:

• What is the atomic arrangement in the thin film and how does it differ from bulk materials?
• Which physical parameters dictate the equilibrium thickness of the thin film and how can the be exploited?
• How is the electronic structure of the thin layer affected by its novel atomic structure? Can the electronic structure be calculated with methods similar to those used for bulk materials? What kinds of experiments are required to resolve electronic structure at such small scales?
• What kinds of optical (or photonic) properties will such films possess---how can they be predicted or characterized by experiment.
• How does the chemistry of the bulk phase and crystallographic misorientation of the ceramic affect film behavior?

Such questions are seldom independent. Their investigation requires combinations of specialized skills and equipment that are not found in a single institution.

High Resolution Transmission Electron Microscope (HRTEM) Image of a Grain Boundary Film in Strontium-Titinate This image was taken by the Stuttgart group and it is an example of many such determinations of the atomic structure adjacent to a grain boundary phase. The University of Pennsylvania group has performed high spatial resolution vacuum ultraviolet (VUV) spectra characterization of such boundaries. Their results are described below. This is a Sigma 13 grain boundary in Fe-doped SrTiO3 containing an disordered intergranular film of 0.6-0.9nm thickness. For this boundary the Stuttgart TEM group has recorded valence EELS (so-called low-loss) spectra, from which all optical properties can be determined on a local scale. This work was done in co-operation with Roger H. French (UPenn&Dupont). Results were compared to VUV spectra of bulk SrTiO3.

High Resolution Transmission Eletron Microscope Image Demonstraing Boundary Films in Silicon Nitride This is a high resolution transmission electron microscopy (HRTEM) image of a Si3N4 ceramic doped with 2wt% Y2O3. The micrographs shows anintergranular film of 1.0+/-0.2nm, determined by both manually definingthe last crystalline lattice plane, and analysing integrated intensityline profiles perpendicular to the film.

Scanning Tunnelling Electron Microscope Image of Silicon Nitride Grain Boundary Film This figure contains 2 bright field STEM images of an Si3N4 ceramic doped with Y2O3 and MgO, taken at different magnifications. The upper left shows a typical part of the materials microstructure containing Si3N4 grains surrounded by glassy material. The upper right image shows a grain boundary film appearing as dark contrast between two crystalline grains of pure Si3N4. In the left grain some lattice fringes parallel to the grain boundary are visible. The image on the lower right is a HAADF micrograph taken from exactly the same specimen area. Since the intensity in this image is roughly proportional to Z**2, one can clearly sea the interface film containing Y. On the lower left the intensity line profile extracted from this image shows the Y segregation to the IGF as well.

Electron energy-loss near-edge Spectroscopy of a Silicon Nitride Grain Boundary Film Electron energy-loss near-edge structure linescans were recorded from pure Si3N4 (containing 4%SiO2) materials. The upper image shows a background stripped spectrum image of one linescan. In the lower image, two spectra of the Si L-edge were extracted from the linescan, one stamming from bulk Si3N4 and one from the intergranular film. The interfacial spectrum shows an additional peak, which is absent in the bulk. This feature indicates additional unoccupied electron states for Si within the grain boundary. From this, a more detailed analysis of the electronic structure and therefore the bonding within the IGF becomes possible. Similar observations were made for the N K-edge and also for other (doped) Si3N4 materials.

Illustration of High Spatial Resolution Optical Spectra of a Grain Boundary This valence EELS spectra of a grain boundary in Strontium Titanate was obtained by the University of Pennsylvania group and illustrates the change in energy of electron interband transitions as a function of position across a grain boundary film in strontium titinate. The differing energy-band structure of the grain boundary has an important consequence on the stability of such films. One can integrate the optical frequencies at each position and obtain a spatially resolved index of refraction. Such an inhomogeneous index of refraction produces dipole-dipole forces across the boundary. These forces are known as Hamaker forces. The spatially resolved indices of refraction that result from such scans across two different grain boundary misorientations are illustrated below. Models for Hamaker forces for such inhomogeneous films have important implications in thin biological films. These models are being developed as part of a collaboration between the University of Pennsylvania group and researchers at NIH

Model Results for Hamaker Forces from VUV Optical Spectroscopy Data The heterogeneous index of refraction model of the NIH/University of Pennsylvania group takes spectroscopy data, like that above, and gives predictions of the cumulative Hamaker forces across the boundary captured in a Hamaker constant. Ultimately, the film thickness is determined by a compromise between forces, such as, structural and osmotic, that tend to result in thick boundaries and the Hamaker forces that tend to thin boundaries. Such results will be of fundamental importance in the prediction and engineering of boundary thickness.

Interband Transition Energies for Ten Distinct Phases in the Y-Si-O-Al-N system These are the results of ab-initio calculations performed by the University Missouri-Kansas City group. This group combines the positions of atoms with electronic structure data to produce band structure predictions for particular atomic arrangements. It is anticipated that atomic positions from molecular dynamics calculations from the Rutgers group and Monte Carlo calculations from the Oxford group might be used as input to these ab-initio models. Subsequently, these results will be compared to VUV spectroscopy results like those performed by the University of Pennsylvania group. Comparisons with the crystal structures simulated above should give insight into the atomic and electronic neighborhood of ions in the thin films.

Continuum Models of Chemistry and Structure in Thin Films These are intermediate results from continuum thermodynamic predictions performed by the MIT group of the chemical compositions and measures of the atomic structural environment in grain boundary thin films. On the left, predictions of the composition fields for Si, Ca, N, and O, as a function of distance from the center of the film. The model also includes the boundary disorder as a spatially independent field. These results are qualitatively consistent with the coarse-grained measures resulting from the molecular dynamics calculations performed by the Rutgers group. On the right, the results show non-uniform Ca distribution in the film. Charged domains are also observed. These charged domains would have novel electrical properties and have led the MIT group to develop theoretical models for the development and stabilization of such domains.

Lanthanum-based Oxynitride Glass Saturated with Nitrogen. The processing route of oxynitride glasses (RE-Si-Mg-O-N with different Rare-Earth elements) was implemented in order to allow metling under high SiO partial pressure, despite the high reducing atmosphere of graphite heated furnaces. Doing so, the thermodynamical equilibrium is garanteed and Si3N4 does not react with SiO2, so that nitrogen can be fully incorporated into the glass network. The picture shows an example for the 5 different glasses processed (with La, Lu and Y as rare-earth element and different N-contents). The glasses are transparent and homogeneous. Those samples with diameter over 1 cm were polished to an optical quality on both sides. N-saturated RE-based glasses sample are are the bulk material corresponding to the intergranular amorphous phase of silicon nitride. Using this samples, the properties and structure of the bulk material can be measured, as will be done by the U-Penn group, and then compared to those of the material as thin film.

Higly pure Si3N4 sample sinter-HIPed with 17 vol-% SiO2. Using the technology developed for the processing of oxynitride glasses, Si3N4 - free of sintering additives - was sintered to closed porosity. With a post-HIPing treatment, densities of up to 98 % of the theoretical density can be achieved. Those materials are free of impurities, compared to samples densified with a capsule-HIP treatment, where borosilicate glasses are used. The intergranular film has the simplest composition achievable and only contains Si, O and N. This represents the first system to be modelled from all the groups (MD, Phase Field). Measurements from TEM groups (e.g. EELS by Stuttgart group or RDF by Oxford group) will provide comparison with - and implementation possibilities of - calculations.

Scanning Electron Microscope image of Si3N4 Microstructure First SEM micrograph of a Si3N4 samples sinter-HIPed with 17 vol-% SiO2.