AVS 52nd International Symposium
    Applied Surface Science Monday Sessions
       Session AS-MoA

Paper AS-MoA4
Angle Resolved XPS for Characterization and Metrology of Ultra-Thin Silicon Oxynitride Films

Monday, October 31, 2005, 3:00 pm, Room 206

Session: Electron Spectroscopies
Presenter: C.R. Brundle, C. R. Brundle & Associates
Authors: C.R. Brundle, C. R. Brundle & Associates
G. Conti, Applied Materials, Inc.
Y. Uritsky, Applied Materials, Inc.
P. Mack, Thermo Electron Inc., UK
Correspondent: Click to Email
The analytical tool set used in the wafer processing industry for metrology of thin films is changing because the films are becoming so thin and more complex. One of the new contenders for both characterization (the kind of detailed information needed in development, trouble shooting, and failure analysis) and metrology (rapid measurements of specific parameters on large numbers of samples) is XPS. Whereas many of the techniques used in the past become less effective as films become ultra-thin (0 to 50A range), XPS becomes most effective in this range. For the high k dielectric material, SiON, as used in the current generation of gate oxide production (10 to 30A), Angle Resolved XPS can provide very precise film thickness and at the same time, in principle, provide a depth distribution of the nitrogen dose using a constrained model to fit the data. Since the model fit to the data can never be unique the approach works best for two extremes: 1) when there are really gross differences in distributions between films and it is this qualitative difference one is looking for and 2) when there are multiple measurements available for films made under a given processing condition that give the same profile fit, and then multiple measurements under a different processing condition which give a consistently different fit (even if small), signifying that a change in profile has occurred. We present examples of both situations from current industry samples. The former is more of a film/process development issue and the latter is more of a metrology issue. It does seem possible to tell that small changes in distribution have occurred, even if one cannot fully quantify those changes. An interesting point is that in none of the many (hundreds)of samples studied, is a spike in the N concentration observed at the surface (first 2A) by ARXPS, unlike in some reports using TOFSIMS on similar samples.