Reducing powder bed layer defects in slurry-based three dimensional printing

Abstract

Slurry-based Three Dimensional Printing is being used to create ceramic parts directly from CAD files. Discrete slurry layers are deposited, into which a binder material is selectively ink-jet printed. This process is repeated until the last layer of the powder bed is deposited. Afterwards, the powder bed is re-dispersed in water, leaving behind the printed green part. The green part is then sintered to full density. This thesis focuses on methods of depositing the slurry layers. Currently, slurry layers are deposited by nozzle rastering. In this approach, a nozzle mounted to an x-y linear positioning system deposits adjacent discrete lines of slurry on a powder bed. Powder beds produced by nozzle rastering contain defects that occur between line and layer interfaces. The top surface has an inherent roughness due to the peaks and valleys between discrete lines. Line merging is a new method of slurry layer deposition that has been developed in an effort to eliminate inter-line defects, improve layer surface finish, and increase throughput This new technique has been used to rapidly produced slurry layers containing fewer internal defects and smooth surface finishes. Line merging occurs when adjacent lines of slurry are deposited in rapid succession such that they merge together prior to slip casting. Line merging differs from nozzle rastering in two ways: lines are deposited in only one direction (during the return pass the nozzle is put into a catch position), and the cycle time between depositing lines is reduced from approximately I second to as little as 0.1 second. A model was developed in an effort to identify the conditions required to achieve successful line merging, while avoiding layer defects such as bubbling and irregular surface finish caused by slurry migration. This model emphasized three relationships: the ratio of cycle time for line deposition to slip casting time for a slurry layer, the ratio of line width to line spacing, and the inverse of the width of the wet slurry zone where lines have merged prior to slip casting. A 3-D plot was constructed relating an objective function comprised of the three relationships to the control parameters (flow rate divided by nozzle velocity and cycle time). A plot for each alumina slurry solids loading was used to guide experiments. These experiments supported the model, though some relationships were proved more accurate than others. The model was ultimately used to target the ideal line merging conditions that were used to produced a 60 layer alumina powder bed out of 50 micron thick layers of 18 vol% alumina slurry. This powder bed exhibited excellent surface finish, with a maximum variation of 11 microns peak to valley. SEM analysis of cross-sections revealed that internal defects between deposited lines, previously seen with nozzle rastering, had been eliminated. Micro-bubbles along the interface between layers persisted, however. Follow-up SEM analysis of a 5 layer powder bed built with 22 vol% alumina slurry revealed no inter-line or inter-layer defects.