Property enhancement of 3D-printed alumina ceramics using vacuum infiltration

https://doi.org/10.1016/j.jmatprotec.2014.01.019Get rights and content

Highlights

  • Vacuum infiltration process was successfully utilized to greatly enhance mechanical and physical properties of alumina ceramics made by powder-bed inkjet 3D-printing process.

  • A model was provided to predict density enhancement of 3D printed ceramics subjected to infiltration and was validated based on experimental data.

  • The relative sintered density of the printed sphere with 1 cm diameter improved from 37% to 86% by infiltration process.

  • 4-pt bending strength of infiltrated ceramic bars was enhanced by 15 times.

  • An infiltrant solids loading of 35–40 vol% ensures full penetration in a 1 cm thick component.

Abstract

A combination of powder-bed inkjet 3D-printing and vacuum infiltration was utilized for manufacturing alumina ceramic components with enhanced mechanical and physical properties. 3D-printed parts were vacuum infiltrated with highly concentrated slurries to improve the green density followed by sintering for densification. The effect of infiltrant solid loading on the physical and mechanical properties of the produced parts was investigated using Archimedes’ method, mercury porosimetry, microCT, 4-point bend test and surface profilometery. In case of parts with small size below 1.5 cm thickness, higher slurry concentration leads to improved density, lowered porosity, enhanced mechanical properties and improved surface finish. However, due to high viscosity of concentrated slurries and incomplete infiltration, thicker parts exhibit lower density at higher infiltrant concentration. The present work provides an effective approach for enhancing properties of inkjet 3D-printed ceramics.

Introduction

Inkjet 3D-printing is one of the most traditional forms of powder-based additive manufacturing technologies which was invented in MIT (Cima et al., 1989) and allows a physical object to be built directly from computer aided design data. In this technique, the powder is spread by a rotating roller onto a build platform inside a build chamber. By means of ink-jet printing technology, a print-head cartridge, consisting an array of piezoelectric nozzles, rasters a liquid jet across the layer of the powder and deposits binder or solvent droplets in those locations defined by the 2D cross-section of the object. Subsequently, the build platform goes downward by one layer thickness and a new layer of powder is spread and the process continues until the whole object is built inside a powder bed. Inkjet 3D-printing is considered a low cost additive manufacturing technology and does not require the use of sophisticated technology. The printing speed is also comparable or better than most of the other methods. Furthermore, it does not require any external support during the printing as it is self-supported by the unbound powder. Thus, it allows the manufacturing process to be simplified and cost efficient.

To manufacture ceramic materials using inkjet 3D-printing, the common method is to mix ceramic powder with a binder which will be activated by adding a solvent on it through a jetting process. This method has been well explained by Utela et al. (2008). One of the critical challenges for this process is the flowability and spreadability of powders. Using powders with large particle sizes will improve the flowability but will compromise the sinterability and densification behavior of the powder after printing. On the other hand, using very fine particle size (<1 μm) will cause severe agglomeration and poor flowability.

One main drawback of inkjet 3D-printing is that the printed object has considerable amount of porosity due to random agglomeration, high friction among particles and lack of any external force to compress the powder and provide better packing. Consequently, 3D printed structures often do not offer the green density required for a full densification in subsequent sintering step. Although spray-dried powders can enhance the tap density of powder to some extent, the resultant packing is not sufficient for densification to above 80% of theoretical density. Suwanprateeb et al. (2010) used spray drying technique and milling to prepare the 3D printing feedstock to make hydroxyapatite parts. The highest density after sintering rarely exceeds 50% of theoretical density of hydroxyapatite. Other post processing techniques rather than sintering is required to improve density. The most common post processing method is infiltration while cold isostatic pressing (CIP) is an alternative. Infiltration of 3D printed ceramics with metallic materials has been demonstrated by Melcher et al. (2006). They printed porous alumina using 3D printing and infiltrate molten copper into the pores using vacuum casting. Dense metal/ceramic composite with very low shrinkage was created using this method. Although the most common infiltrant is molten metal for making ceramic metal composites, in this research we focused only on pure ceramics. In one approach as demonstrated by Yoo et al. (1993) the 3D printed ceramic is infiltrated with a glass material penetrating into the pore channels by capillary effect. They produced fully dense composite ceramics with a low linear shrinkage below 1%. In another approach, the printed part is infiltrated with a ceramic slurry or sol. The process is usually conducted after sintering where the object with higher density and strength is prepared. Lee (2001) had conducted the infiltration technique of alumina sol on 3D-printed parts obtained by Selective Laser Sintering. The density and bending strength of the sample improved up to 100%. Similar experiments had been conducted by Vogt et al. (2010) to improve the properties of ceramic foams by vacuum infiltration. By adjusting the slurry characteristics, they managed to improve the strength up to 388%. This improvement in properties is due to the vacuum infiltration refilling the hollow struts.

In this work, a slurry vacuum infiltration process which is more effective than normal infiltration, was utilized for density enhancement of alumina parts made by inkjet 3D-printing. A highly solids loaded alumina slurry was forced into a 3D-printed alumina skeleton which subsequently subjected to a firing cycle. The physical and mechanical properties of the produced alumina ceramics were evaluated. The slurry parameters which affect the penetration depth and final densification of the component were investigated.

Section snippets

Raw materials

High purity alumina powder (Sumitomo AKP30, Japan) with a nominal particle size of 0.32 μm was used to produce alumina ceramics. Poly vinyl alcohol (PVA, Nippon, Japan, Mw: 17,000 g/mol) was used as a binder. Dolapix CE64 (Zschimmer & Schwarz Chemie GmbH, Germany) was used as a dispersant for preparation of alumina slurries.

Powder preparation for inkjet 3D-printing

Powder for inkjet 3D-printing process with high flow ability and suitable agglomerate size (<50 μm) was prepared using spray drying technique. For this purpose, alumina aqueous

Results and discussion

Fig. 2 shows the morphology of the spray-dried alumina powders used for inkjet 3D-printing process. It can be seen that the powders comprise spherical agglomerates with a nominal size of 20–50 μm and each agglomerate contains submicron sized particles. The spray-dried powder shows much better flowability and spread ability than raw powder, which prevents the formation of large defects in the part during printing process and dramatically improves the handling strength of part during depowdering.

Conclusions

A combination of powder-bed inkjet 3D-printing and vacuum infiltration process was utilized for producing alumina parts with higher density and improved mechanical properties. A suitable powder system based on spray drying process and dry mixing was applied for preparation of a feedstock with sufficient flow ability for printing process. Since the packing of the green parts after printing is quite low, vacuum infiltration process was utilized to impregnate the 3D-printed parts with highly solid

References (7)

There are more references available in the full text version of this article.

Cited by (0)

View full text