Graded (Ti_1-x Al_x)N coatings were prepared by reactive magnetron co-sputtering of Ti and Al on silicon and glass substrates in an Ar+N₂ atmosphere. The N₂ content was kept constant in the gas mixture and the composition of the (Ti_1-x Al_x)N film was controlled by the power applied to the targets.

The composition of the graded films was determined both by Auger electron spectroscopy (AES) depth profiling and by energy dispersive spectroscopy (EDS) of their cross sections. Phase determination was performed by x-ray diffraction (XRD) and by selected area electron diffraction (SAED). Transmission electron microscopy (TEM) of the cross sections was used to determine the microstructure of the graded films. Interfaces between the films and substrates and within the films were investigated by high resolution TEM (HRTEM).

AES depth profiling and EDS from the cross sections confirmed the formation of (Ti_1-x Al_x)N films with graded composition. XRD and SAED showed that only the B1-NaCl phase was formed as typical for Ti-based nitrides. TEM observations of the cross sections revealed that the films had a columnar microstructure. TEM and HRTEM inspections have shown that continuos graded films with no abrupt interfaces could be deposited by a proper choice of the process parameters.

Hydrodenitrogenation (HDN), hidrodesulfurization (HDS) and hydrodeoxigenation (HDO) are the most important catalytic processes in converting crude oil and coal derived liquids into clean-burning fuels. Nickel and/or Cobalt promoted Mo and/or W sulfide catalysts are commonly used for these reactions. While Mo and W sulfide based catalysts are effective in meeting current industrial processing objectives, more stringent environmental pollution limits will require the development of new, more active and selective catalysts.

Transition metal carbides have shown great promise for use as commercial HDN and HDS catalysts. The electronic, structural and catalytic properties of early transition metals have been the subject of many experimental and theoretical investigations since these carbides often demonstrate unique catalytic advantages over their parent metals in activity, selectivity and resistance to poisoning.

Recent results in the field of catalysis indicate that part of the activity of HDS catalysts may be due to the presence of carbides, since these compounds also present substantial catalytic activity in hydrotreating reactions, especially the carbides from transition elements such as molybdenum and tungsten.

In order to identify the carbon species in real samples, we firsts decided to perform a complete comparative EELS study of ten commercial carbides. We found that at the low energy window (0 - 80 eV), the most prominent plasmon peak and other high order edge transitions make spectrum shape sufficiently distinct to allow chemical identification possible. These differences among the various carbides are more accentuated at the energy window where the carbon K-edge transition is located (250 - 400 eV). The differences in this case are more striking because the chemical state is different in each structure. Thus, it is possible to asses the chemical state of each structure and it is even possible to differentiate between two chromium carbides.

The performance of advanced thin-film devices often depends cruci ally on their structures. Therefore, precise control of growth process and c haracterization of film thickness, surface and interface roughness with high resolution has become an important task. Grazing incidence X-ray reflectivit y is a powerful technique for investigating surfaces and multilayers. GIXR o ffers high spatial resolution, which can be as good as 0.1 nm with the measu rement of roughness and thickness. Its high penetration and nondestructive c apability are very suitable to probe buried interfaces or to study real-time processes.

In this work, several different material systems, such as Ta₂O₅/SiO₂, Co/Cu, and SiN/Si multilayers, were studied by GIXR. The thickness determined by GIXR is found in good agreement with that by TEM . The different interface roughness feature and evolution in multilayers, th e effect of substrate roughness and annealing on the interfacial structures are successively characterized. The details will be presented.

According the mixing - roughness - information depth (MRI) model¹ the depth resolution function necessary for the reconstruction of compositional depth distribution from sputtering profile can be described by three physically well defined parameters: atomic mixing, surface roughness and information depth. Typical test structures for the MRI model are multilayers with well defined interfaces. The surface roughness is composed of the original roughness of an interface and the roughness induced by ion sputtering. In this study, we try to separate the original and induced roughness by using reference SiO₂ / Ta₂O₅ multilayers samples. AES depth profiles were acquired using the low and high energy peaks of Ta and Si and the oxygen peak, which shows a shift of 6.9 eV from Si to Ta oxide layers. All peaks were used for MRI analysis. The sputter depth profiles were fitted with the MRI model in two ways, (a) without taken in to account the original roughness and (b) with the assumption that the effective roughness R_ef is equal to

R_ef = (R_or² + R_ind² )^1/2

where R_or and R_ind are the original roughness of the interface and the roughness induced by ion sputtering, respectively . The original roughness R_or of the interfaces was measured by means of grazing incidence x-ray reflectivity (GIXR). R_or varies from interface to interface which is also recognized in depth profile. The values of R_or are in the range between 0.23 and 0.68 nm. The values of R_ind were found by MRI fitting and compared with AFM measurements. Both the GIXR and MRI methods for determination of the roughness support each other and allow a more precise determination of the depth resolution function.

¹ S. Hofmann, Surf. Interface Anal. 21, 673 (1994).