Mechanisms for agglomeration and deagglomeration following oblique collisions of wet particles

Abstract

Previous studies on wetted, particle-particle collisions have been limited to head-on collisions, but in many-particle flows, collisions are inherently oblique. In this work, we explore such oblique collisions experimentally and theoretically. Whereas in normal collisions particles rebound only due to solid deformation, we observe in oblique collisions a new outcome where the particles initially form a rotating doublet and then deagglomerate at a later time due to so-called centrifugal forces. Surprisingly, we discover the essential role of capillary forces in oblique collisions even when the capillary number (viscous over capillary forces) is high. This recognition leads to the introduction of a dimensionless number, the centrifugal number (centrifugal over capillary forces), which together with the previously established Stokes number characterizes the regime map of outcomes. Unexpectedly, we observe a normal restitution coefficient greater than unity at large impact angles, the mechanism for which may also be observed in other agglomerating systems.

Agglomeration and de-agglomeration of rotating wet doublets

“…Two of the outcomes were also observed in previous experimental works involving wetted normal particle-particle and oblique particle-wall systems, namely stick (S, figure 5a) in which the striker particle forms an agglomerate with the target object, and bounce (B, figure 5c) in which the striker collides with the target and essentially immediately separates (Kantak . The third outcome is only observed in oblique particle-particle collisions and is shown in figure 5(b) (Donahue et al 2012). Here, the two particles collide and initially stick together, and then at a later time after the agglomerate has rotated through a significant angle, the two particles de-agglomerate due to centrifugal forces.…”

“…Except for one instance, string tension generally does not affect the observed trends in e w,n and θ R for de-agglomerating collisions. However, for a given impact angle, the value of St n that demarcates agglomeration versus de-agglomeration, St * (θ 0 ), is largely influenced by string tension, since without string tension all oblique collisions are predicted to de-agglomerate (as also shown in Donahue et al 2012); therefore, in this work the trends of St * (θ 0 ) are not analysed. (In figure 7, with string tension,…”

“…Recently, we observed a new outcome for oblique, particle-particle collisions in which two colliding particles initially form an C. M. Donahue, W. M. Brewer, R. H. Davis and C. M. Hrenya agglomerate and then later de-agglomerate due to centrifugal forces (Donahue et al 2012). In that work, a corresponding theory was developed that agrees well with the regime map of outcomes for impact angle versus St n , and both experiments and predictions show that the new outcome occurs at large impact angles for St n < St * n .…”

“…Surprisingly, the trends depend on whether a de-agglomerating collision occurs for St n > St * n or St n < St * n . Furthermore, the theory established in Donahue et al (2012) is able to predict the trends and provides further insight into the physical mechanisms responsible for particle reversal and de-agglomeration.…”

Effects of aggregation on the kinetic properties of particles in fluidised bed granulation

“…This means that any relative collision energy between the colliding particles less than that would succeed for aggregation [53]. It is worth mentioning that this energy can be explored to include the rotation [84,85] of the aggregating particles. However, the calculation of θ* is not the scope of this article so will not be sought.…”