| The elementary process of polymer melt devolatilization (DV) has been investigated experimentally as well as theoretically. The rolling pool present in all rotating machinery used in DV is considered as a locus of polymer DV. A batch-mode process was developed in our lab, using a drum-rotating device, that permits independent control of all pertinent DV parameters. Parametric studies include exposure time ((lamda)), total pressure (P), agitation rate (N(,R)), initial volatile concentration (W(,O)), volatile vapor pressure (P(,s)), surface-to-volume ratio, and addition of inert substances. Volatile levels in all samples were analyzed utilizing a headspace gas chromotagraph method specially designed for this purpose. Important parameters for DV enhancement are total pressure, initial volatile concentration, and agitation rate. Then, a new advanced tube-rotating apparatus was built to further explore the detailed effects of individual and/or grouped parameters, such as SH(,O) (K(,w)W(,O)-P), chemical driving force, and mechanical driving force (P(,me)-P), where SH is superheat and K(,w) is Henry's law constant.; It has been found that modest SH(,o) along with agitation are able to cause vacuum foaming. A bubble formation theory was thus developed. The cavity model based upon pre-existing gas cavities leads to one plausible formation mechanism to explain foam-enhanced DV. The theory appears to agree with observations. Besides, since variation of internal pressure with variation of N(,R) can vary bubble expansion resistance, a (psi)-(omega) fluid mechanical model was established to study velocity profile and pressure profile inside the rolling pool. The computational solution seems consistent with observations. In an attempt to explain bubble growth phenomena, a cell model was adopted and the primary role of nucleation rate (N(,b)/V(,p)) in foaming separation has been confirmed. We have also developed an exponential model to demonstrate the overall separation in the case of entrained-foaming DV. The exponential curves show reasonably good agreement with experimental results. |