Effects of Varying Surface Film Thickness On Particle Adhesion In Semiconductor Material-Based Systems

Effects of Varying Surface Film Thickness on Particle Adhesion
in Semiconductor Material-based Systems

Abstract

The effects of surface film (coating) thickness, d, on the adhesion of a particle to a surface are investigated.  The model system in this work includes a pre-manufactured silicon nitride probe (initial radius of curvature ~15 nm) and a silicon substrate onto which gallium nitride films have been deposited via atomic layer deposition.  The gallium nitride films have varying thicknesses (0, 5, 11, 12, 16, and 80 nm).

We develop theory supporting the hypothesis that the coating material will appear transparent to the particle if d is small in comparison to the particle-substrate separation distance, D.  However, as d increases a transition in the particle-substrate interaction will take place.  This implies the existence of a critical coating thickness, d*, beyond which the adhesion behavior of the coating material will be equivalent to bulk (see Fig. 1).  We show d* is a simple function of D and a parameter termed the permissible fractional error, ε.  Taking D = 0.4 nm at contact, for ε = 0.05 (i.e. 5% permissible fractional error), d* ≈ 28 nm.

The adhesion between the model system particle and substrate is measured using contact mode atomic force microscopy (AFM).  Experimental results confirm that for small (i.e. 5, 11, and 12 nm) d, the interaction between the silicon nitride probe and underlying silicon is relatively unaffected by the presence of the gallium nitride film; for large (i.e. 80 nm) d, the interaction parallels that between silicon nitride and bulk (≥ 5 µm) gallium nitride.  The d* for this model system is approximately 30 nm, which is in good agreement with our theoretical predictions.

These developments were then examined on the industrial scale by conducting Particle Removal Efficiency (PRE) experiments on six 300-mm diameter SiO₂-on-silicon wafers, onto which varying thicknesses (5, 10, 15, and 20 nm) of titanium nitride had been deposited.  The challenge particles were dry-deposited SiO₂ particles having size distribution peaks at 55 and 110 nm in diameter.  Wafers were cleaned using a commercially available argon/nitrogen cryogenic aerosol process.  Results from the PRE experiments (see Fig. 2) revealed that SiO₂ particle removal efficiency improves with TiN film thickness, implying TiN shields or screens the adhesion of the SiO₂ contaminants to the underlying layer of SiO₂.  Additionally, PRE improves with particle size; PREs for the 110-nm challenge particles were 1.5×–4× times those for the 55-nm challenge particles.  Finally, as TiN film thickness approached d*, changes in PRE became more drastic.

Figure 1:  variation in AFM pull-off force with GaN surface film thickness.

Figure 2:  variation in PRE with TiN surface film thickness.

References

[1] J. N. Israelachvili. (1992). Intermolecular and Surface Forces (2^nd Ed.). Boston: Academic Press.

[2] R. P. Jaiswal. (2008). Adhesion between particles and nano-structured films. (Doctoral dissertation). West Lafayette: Purdue University Press.