Starve Feeding in Screw Extruders: A Review

Starve‐fed polymer extrusion has become an important strategy for improving process control, energy efficiency, and product quality, but it also introduces partially filled flow regimes that are not captured by classical flood‐fed concepts. This review provides a critical assessment of starve‐fed operation in single‐screw and corotating twin‐screw extruders, structured by extruder type and process region. Key experimental techniques, including screw pull‐out/quench analysis, window‐based and optical visualization, residence time distribution (RTD) measurements, mixing assessments, ultrasonic sensing, pressure profiling, and specific mechanical energy (SME) monitoring, are reviewed with emphasis on how starve feeding modifies melting mechanisms, fill level, RTD, and energy consumption. The experimental evidence base is particularly strong for starve‐fed single‐screw extrusion, where pull‐out and RTD studies clearly demonstrate a transition from contiguous solid‐bed melting to mixed conductive/dispersed melting; corresponding data for starve‐fed twin‐screw extruders remain comparatively scarce and are often limited to local visualization and RTD in selected elements. Modeling approaches are then summarized, from early analytical and empirical models to modern global feed‐to‐die descriptions utilized to partially filled operation. Their capabilities and limitations are discussed separately for single‐ and twin‐screw machines, highlighting where starve‐fed conditions can be treated by adapted flood‐fed models and where new formulations are required. Finally, advanced numerical techniques, computational fluid dynamics (CFD), discrete element method (DEM), and smoothed particle hydrodynamics (SPH), are reviewed with respect to extruder type and phase: DEM for granular solids conveying in the feed region, CFD for free‐surface melt flow in partially filled single screws and selected twin‐screw sections, and SPH for resolving starved and intermeshing flows with explicit free surfaces. The achievements and current limitations of these methods are identified, along with future research needs in improved modeling of partially filled regions, tighter coupling between DEM and CFD/SPH, real‐time process monitoring, and scale‐up strategies for industrial starve‐fed extrusion.