AVS 45th International Symposium
    Thin Films Division Friday Sessions
       Session TF-FrM

Paper TF-FrM9
A Novel Method For Determining Kinetic Rate Expressions For CVD Processes Using a Combination of Step Coverage Measurements and Computer Simulation

Friday, November 6, 1998, 11:00 am, Room 310

Session: Thin Film Deposition from Chemical Precursors
Presenter: G. Ramanath, Novellus Systems
Authors: E.J. McInerney, Novellus Systems
G. Ramanath, Novellus Systems
D.C. Smith, Novellus Systems
Correspondent: Click to Email
There is an increasing need for understanding reaction kinetics and mechanisms during chemical vapor deposition (CVD) to accurately predict deposition rate, uniformity, step coverage, and reactor throughput. We present a novel method for determining reliable kinetic rate expressions by a combination of transport modeling and experimental measurements of step coverage and planar deposition rate. We demonstrate the method by applying it to W CVD during H@sub 2@ reduction of WF@sub 6@ to elucidate -- for the first time -- two distinct kinetic mechanisms operating in different WF@sub 6@ flow regimes. Rate expressions deduced from deposition rate measurements on planar geometries often fail to accurately predict step coverage due to incorrect assumptions of near-surface concentration N@sub s@ of reactants. As direct measurements are difficult, modeling species transport in the reactor is necessary to estimate N@sub s@. However, this method by itself is inadequate because small errors in the model can lead to large errors in the rate expression. In our method, we obviate the limitations of traditional modeling by utilizing experimentally measured step coverage values to model species transport within topological features in addition to that in the reactor. In this way N@sub s@ is calculated self-consistently, allowing the determination of the reaction order n and kinetic mechanisms. Despite numerous studies of the H@sub 2@-WF@sub 6@ reaction over the last 30 years, the order of WF@sub 6@ concentration dependence on W deposition rate has been unresolved between n=0, 1/6, and 1. Our results reveal two distinct mechanisms: at high WF@sub 6@ flows the rate is HF desorption limited (n=1/6), while at low flows WF@sub 6@ adsorption is the limiting step (n=1).