AVS 49th International Symposium
    Thin Films Thursday Sessions
       Session TF-ThA

Paper TF-ThA10
Characterization of Ultra-Thin Films using Angle Resolved XPS and Maximum Entropy Methods

Thursday, November 7, 2002, 5:00 pm, Room C-101

Session: Ultra Thin Films
Presenter: R. White, Thermo VG Scientific, UK
Authors: R. White, Thermo VG Scientific, UK
R. Champaneria, Thermo VG Scientific, UK
J. Wolstenholme, Thermo VG Scientific, UK
P. Mack, Thermo VG Scientific, UK
Correspondent: Click to Email
The Thermo VG Scientific Theta Probe has the unique ability to collect angle resolved XPS (ARXPS) spectra in parallel without the need to tilt the sample. The collection of ARXPS data (free from artifacts associated with sample tilting) is now routine for samples ranging in size from 20 µm to 300 mm. This has made it necessary to implement a consistent method for the generation of depth profiles from ARXPS data. The conversion of ARXPS data into depth profiles is problematic as no single transformation from ARXPS data to depth dependent concentrations exists. As real data contain noise, a weighted sum squares of error between measured and calculated data is not sufficient to determine accurate depth profiles. A usable solution can be found by the inclusion of an entropy term that allows a profile to be generated with the minimum amount of structure. Characterisation of ultrathin high K dielectric overlayers on SiO@sub 2@ is one application where ARXPS shows great promise, potentially yielding thickness, distribution, dosage and chemical state information. A maximum entropy approach based on the methods developed by Smith and Livesey has been adopted in order to interpret Theta Probe data. This approach may be thought of as a least squares fitting of the data, with an in-built 'smoothness' parameter to prevent overfitting to noise in the data, for example. Maximum entropy calculations are not restricted to small numbers of chemical state components. This makes this method particularly suitable for ARXPS where low concentration and multi chemical state components need to be profiled. This paper describes the method and results from various systems where this approach has been adopted.