Failures in disk drives are frequently induced by particulate contamination present in the head/disk assembly (HDA). The sources and characteristics of debris vary with component material within the HDA. The inherently contaminating nature of manufacturing component parts requires that those parts be carefully cleaned and that their cleanliness be evaluated against a specified standard. Ultrasonic cleaning of substrate surfaces, and turbidity/liquid-particle-count (LPC) measurement of extracted-contaminant concentration and size distribution are popular techniques in this regard.
Two factors can be used to determine the suitability of ultrasonic surface cleaning. The first is defined here as the maximum cleaning potential, that is, how clean the surface can get when subjected to a multistage-extraction process. And the second concerns the first-stage cleanability of the surface, that is, what fraction of the total cleaning action occurs in the first ultrasonic stage. These two fundamental cleaning parameters described here are closely linked to the nature of adhesive interaction between the contaminant particle, the substrate surface, and any intervening liquid. A predictive procedure to rank various materials with respect to their maximum cleaning potential and their first-stage cleanability is presented. Turbidity/LPC data, obtained at IBM San Jose laboratories for several metallic and polymeric parts, are used to develop useful correlations in terms of relevant physical properties of these surfaces. The implications of this predictive model for future design and optimization of ultrasonic cleaning/monitoring processes are also discussed.