With the addition of dry ice to the flowing gas stream, a new process is available for particle removal. This process is schematically shown above where the particle on the surface is impacted by an impinging dry ice snow. The impact implies momentum transfer between the dry ice and surface contaminant and this momentum transfer can overcome the surface adhesive forces. Once liberated from the surface, the contaminants are easily carried away with the high velocity gas stream. Particle removal efficiencies do not decrease for smaller particulate sizes as in the case for the aerodynamic drag force. Further, the ability to remove surface contamination by momentum transfer appears to be independent of surface contaminant size.
The removal of organic residues relies on two different mechanism, one involves the presence of liquid carbon dioxide at the surface being cleaned. Liquid carbon dioxide is an excellent solvent for hydrocarbons and other nonpolar substances. Please note, liquid carbon dioxide is thermodynamically unstable at atmospheric pressure and room temperature but is present in the snow cleaning process on the surface. An impacting snow particle on a surface is shown schematically below. During the short impact time, the pressure increases at the snow - surface interface and this pressure can easily exceed the yield stress and triple point pressure of the dry ice particle (78 psi). Then, the dry ice particle can liquefy and act as a solvent while in contact with the surface. Contamination would be dissolved by the liquid phase and stay within the "liquid" section. When the particle starts to rebound off the surface, the interfacial pressures decrease and the dry ice particle re-solidifies, carrying the contamination away.
As pointed out by Whitlock, this surface liquid phase may explain why carbon dioxide snow cleaning is a gentle cleaning process. The presence of a liquid phase would limit the pressure that can be imparted to the surface during cleaning. When the snow starts to liquefy at the surface/dry ice interface, continued yielding and plastic flow only tend to increase the size of the contact area liquid phase. Although the solvent action is highly effective, the cleaning process has not achieved removal efficiencies similar to SFCO2. These differences are not clear but may be related to the lower temperature and pressure for the impinging snowflake than for the supercritical phase. The presence of this liquid phase may also assist in particle removal (see below).
M. Hills {1995} explored the removal of various organics compounds using CO2 snow. In her study, she established data supporting the above liquid phase mechanism. Further, data presented established a rule of thumb to determine what compounds can be easily removed. Generally, organics, which are easily absorbed in liquid CO2 are removed; organics with complex or other functional non-hydrocarbon based groups not readily soluble in liquid CO2 are not easily removed. Instead, she suggested a freeze-fracture removal process for these compounds. The freeze - fracture process is responsible for the removal of silicones, fluorinated compounds and pump oils, and other insoluble organics. We have seen many pump oils, Krytex, silicones layers be removed, even ink on glass. Please note that freeze fracture is more effective with a liquid CO2 source than a gas CO2 source.
Polymer layers are not soluble in liquid CO2 and require a freeze fracture process. For adhesives, epoxies, and fluxes, these polymeric layer will remain after cleaning. Here, surface binding to the substrate, substrate properties, and size of the deposit may be determining factors that prevent removal. Testing is required on each sample class and situation.
One topic that needs separate clarification is the removal of deposited films or coatings by carbon dioxide snow cleaning. Films that are deposited properly will remain on the substrate while those films that were deposited poorly onto dirty substrates may be removed. The reason for the inability to remove well-bonded films may be related to the combination of the nonabrasive action of CO2 snow and the inability of the cleaning process to overcome any chemisorbed, covalent, metallic, or ionic bond.
A recent suggestion has been made that there may be another particle removal mechanism. The liquid CO2 phase when in contact briefly with the surface, can act to change the particle adhesion and may lead to particle removal. A good example is to wet your finger and pick up salt, or any powder. A similar process may occur but has not been proven. There is indirect evidence supporting for this concept. Some evidence has been obtained hinting that smaller incident dry ice sizes lead to better removal of smaller particles compared to larger dry ice pieces. For the same velocity, a smaller dry ice particle will have a larger liquid percentage during impact while a larger dry ice particle may not transform to liquid as much and leave behind more particles.
E. Hill{1995} proposed a different mechanism for surface cleaning with the large snowflake, low velocity streams. Here, a thermophoretic process - forces caused by thermal gradients, can remove the particles off the surface. A fundamental aspect of this proposed process was that the snow flake does not contact the surface; instead, as the snowflake approaches the surface, the large thermal gradient causes sublimation on the bottom of the flake. This large release of gas separates the flake from the surface and also creates large shear forces on the surface particle population and can lead to liberation of the surface particles. Next, Hill postulates that thermophoretic forces attract the liberated particle to the cold side of the flake and are taken away for removal.
