2016 Volume 63 Issue 7 Pages 524-529
All-ceramic artificial teeth were produced using a high-speed centrifugal compaction process (HCP) combined with a resin shell-mold made by a 3D printer. Slurries of alumina or zirconia fine powders filled the inside and outside of the mold and then rotated at between 7,000 and 11,500 rpm in a centrifuge (HCP buried compaction method). Using this method, crack-free green compacts were produced. The shell-molds were not deformed or broken because the inner and outer pressures generated during the HCP were quasi-balanced. Two methods for mold-releasing, thermal decomposition and mechanical de-molding by hand, were investigated. Thermal decomposition introduced the critical problem of sintering inhibition. To obtain the final products, the compacts were air sintered after being released from the molds. For alumina, green compacts of high packing density (63 %) were sintered homogenously without considerable deformation. For zirconia, the packing density reached approximately 55 % with a density gradient. The zirconia compacts were sintered inhomogeneously, which resulted in a density gradient and deformation. The density gradient and shape deformation of the sintered compacts are discussed.
Additive manufacturing (AM), also known as 3D printing, has attracted significant attention as a new materials processing technology. In AM, molten- or liquid-state materials are placed on a substrate and solidify immediately, piling up from bottom to top, to form the product shape. Because AM does not require any molds or dies, it is suitable for production of complex-shaped parts in small quantities. As the technology has become more popular, a wider selection of materials, including metals and ceramics, has become available for AM. However, strength and reliability in AM-processed metals and ceramics remains a challenge because of defect introduction during the process. Resin 3D printers, which are a more established technology, can easily produce accurate and flawless products in a shorter time with lower costs than metal and ceramic 3D printers.
The present study aims to produce beautiful, reliable and low cost all-ceramic artificial teeth. Unlike previous studies, which have employed direct AM of ceramics1–3), we used a combination of resin 3D printing and a high-speed centrifugal compaction process (HCP). The HCP employed is powder compaction method using slurry that is suitable to create complex shapes and dense, homogeneous and flawless powder compact4–6). We used a resin 3D printer to make the molds, packed the powder into the molds using the HCP, and then sintered in air furnaces. This process is similar to that used for AM processed ceramics, which also requires a sintering and de-binding process.
We have produced all-ceramic artificial teeth from high purity alumina and yttria-partially-stabilized zirconia. The sintered density, microstructure and dimensional change during the process were evaluated.
Fig. 1 shows the process of preparing all-ceramic artificial teeth. First, resin shell-mold (a) is produced using a 3D printer. A UV-hardening acrylic resin printer (MiiCraft, made in DLAB) was employed. In this type of 3D printer, the resin is hardened using a sharply focused UV beam generated by an LED laser. The printer had a minimum process step of 50 μm. Two types of molds were prepared: one mold was for a front tooth (Fig. 2 (a)) and the other was for back tooth (Fig. 2 (b)). Both molds were shell-type molds, i.e., they had thin walls. Using shell-type molds reduces the process time and cost of the molds. Each mold was set into a resin tube.
Manufacturing process of all-ceramic artificial teeth.
Resin shell mold design, (a) A-type for front tooth and (b) B-type for back tooth.
Alumina and zirconia slurries were prepared by ball milling the respective powders. The alumina powder was made by Taimicron (Taimei Chemical Co., Ltd.)7,8) and had a nominal grain size of 0.14 μm (Table 1 (a)). The zirconia powder was made from a granular powder of 3Y-Zirconia (TZ-3YS, Tosoh Corporation) with primary grain size of 39 nm (Table 1 (b)). Table 2 shows the components of the slurries4–6,9,10). The prepared slurries were poured into whole tubes, i.e., the inside and outside of the shell-mold were immersed in the tubes (Fig. 1 (b)).
| (a) | ||||||||||||
| Al2O3 [%] | impurity [ppm] | Grain size [μm] | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Si | Fe | Na | K | Ca | Mg | Cu | Cr | Mn | U | Th | ||
| > 99.99 | 10 | 8 | 8 | 3 | 3 | 2 | 1 | < 1 | < 1 | < 0.004 | < 0.005 | 0.1 |
| (b) | |||||||
| Zr2O3 [%] | ingredients or impurity [%] | Grain size [nm] | |||||
|---|---|---|---|---|---|---|---|
| Y2O3 | Al2O3 | SiO2 | Fe2O3 | Na2O | Ig-loss | ||
| 93.9 | 5.27 | 0.005 | 0.002 | 0.002 | 0.009 | 0.4 | 39 |
| Materials | Mass% [%] |
|---|---|
| Ceramics (alumina or zirconia) | 100 |
| Dispersant | 1.5 |
| Binder | 0.24 |
| Ultrapure water | 32.3 |
The tubes (two or four tubes were used to balance the centrifugal force) were set in a centrifuge with an effective rotor radius of 120 mm and rotated at between 7,000 and 11,500 rpm for 40 min (the HCP is shown in Fig. 1 (c)). Table 3 shows the rotating speed and centrifugal acceleration generated at bottom of the tubes. During the HCP, the entire shell-mold was buried in the sediment of the powder. Through this “buried compaction”, the outer and inner pressures of shell-mold were quasi-balanced, which allowed an accurate green compact to be obtained without breaking or deforming the shell-mold. After the HCP, the whole tube was set in a dry oven at 313 K for 24 h to dry the green compact.
| Rotating speed of centrifuge [rpm] | Centrifugal acceleration [G] |
|---|---|
| 7,000 | 6,500 |
| 11,500 | 17,700 |
We investigated two methods for mold-releasing: thermal decomposition and mechanical de-molding. To eliminate the resin mold by thermal decomposition4–6,11), the green compact (covered with the mold) was placed in the air furnace and the temperature was gradually increased until the resin mold decomposed at 673 K. This method has several advantages: i) easy handling of green compacts; ii) easier to release the mold compared with mechanical mold-releasing; iii) simpler process because the mold-releasing and sintering are performed using one heating process. Mechanical demolding involved removing the shell-mold by cutting it off using a hot cutter. Fragments of the mold that remain were removed by hand with dust-free gloves.
2.4 SinteringThe alumina and zirconia green compacts were sintered in an air furnace (Figs. 1 (d)–(e)). The heating program used is shown in Fig. 3. The sintering temperature for alumina and zirconia was 1623 K.
Heating program.
The resin shell-molds for the front tooth (Type A) and back tooth (Type B) were made using the 3D printer. A wall thickness of at least 0.4 mm was necessary to make flawless shells. The bottom of each mold was attached to a 2.5 mm wide column to resist the buoyancy force created during buried compaction by the HCP. Small holes at the top and bottom of the shell-molds allowed the slurry to enter the molds.
After the HCP buried compaction and drying, the outer sediment of the shell-mold could be removed easily by hand (Fig. 4 (a)) to revel the green compact covered with the shell-old (Fig. 4 (b)).
Procedure of de-molding process, (a) as taken off from centrifuge, (b) after taking off of outer sediment.
The green compacts covered with the shell-mold (Fig. 4 (b)) were set in the air furnace and the temperature was increased gradually. Images of samples at each temperature are shown in Fig. 5, which shows how complete de-molding was accomplished. Formation of cracks in the green compacts during the process was suppressed by controlling process conditions.
Appearances of the samples at each temperature in thermal de-molding process.
The fragmented shell-mold and the inner green compact are shown in Fig. 6. Because of the thin wall of the shell-mold, the mechanical de-molding was relatively straightforward. Both type A and B compacts were removed without cracking.
Appearance of green compact and released resin shell mold by mechanical de-molding.
The packing density of the alumina compacts that were pre-sintered at 1073 K for 1 h was approximately 63 % and a density gradient was not observed in the compacts. Because a packing density of 63 % coincides with the random close packing density of the spheres12), we concluded that the process used produced uniformly and densely compacted alumina green compacts.
(b) ZirconiaTable 4 shows the densities of the green compact and the sintered compact at various positions within the compact (bottom, middle, and top). The average packing density of the zirconia green compact (approximately 55 %) was lower than that of the alumina compact. Moreover, a packing density gradient was observed; the packing density increased from the top to the bottom of the compact (the density at the lower rotation speed was 0.8 % higher than at the higher rotation speed). Elevating the rotation speed to 11,500 rpm suppressed the density difference.
| Rotating speed [rpm] | Packing density [%] | ||||
|---|---|---|---|---|---|
| At bottom | Middle | At top | Whole green compact | Sintered compact | |
| 7,000 | 54.7 | 54.2 | 53.9 | 54.1 | 96.9 |
| 11,500 | 55.1 | 55.1 | 54.9 | 55.1 | 95.6 |
Table 5 shows the densities of sintered bodies. Only the mechanically de-molded samples reached a sintered density of 99 %. The thermally de-molded samples had a sintered density of approximately 80 %. The cause of such a dramatic difference is not clear. It may be because the result of contamination from the evaporated resin of the shell-mold. Because of this result, only mechanical de-molding was used for subsequent experiments.
| The method of de-molding | Relative density [%] |
|---|---|
| Mechanical de-molding | 99 |
| Thermal de-molding | 80 |
Fig. 7 shows the appearance of the alumina sample sintered at 1623 K using the type A and B molds. Neither product was cracked or broken. Fig. 8 shows the dimensional change during sintering of the samples shown in Fig. 7. In Fig. 8, the dashed line shows the size of the original design (i.e., the inner wall of the shell-mold), the solid line shows the assumed shrinking dimension in the case that a green compact with a density of 63 % was sintered to 100 %, and the plots are the values measured from sintered compact. Although the top part of the A-type tooth has rough bumps on the surface, the contour plots for both sintered compacts coincide with the isotropic shrinkage line (solid line), which indicates that isotropic shrinkage was achieved. Fig. 9 shows the scanning electron microscope (SEM) images of cross sections of the sintered compact (type A) at each part, which revealed that a fine and homogeneous microstructure was obtained throughout the compact.
Appearance of sintered alumina made with, (a) A-type mold and (b) B-type mold.
Dimensional change after sintering of alumina of, (a) A-type mold and (b) upper half of the compact with B-type mold.
SEM images of cross section of sintered alumina (at 7,000 rpm), (a) at bottom, (b) at middle and (c) at top.
Fig. 10 shows the zirconia compacts sintered at 1623 K. The compacts were not cracked or broken, but the relative density rose to only 96 %. Fig. 11 shows the dimensional change after sintering. Insufficient shrinkages were observed in some parts of the compacts, especially at the top half of the compact at lower rotating speed. These results indicate that zirconia was not sintered sufficiently. Although the deformation decreased with higher rotating speeds, some deformation remained. Fig. 12 shows the SEM image of cross sections of the zirconia-sintered compacts. There were significant differences in microstructures at the bottom and top of the sample. At the top of the sample, a number of original powders that were not incorporated during the sintering densification remain.
Appearance of sintered zirconia made with A-type mold.
Dimensional change after sintering of zirconia processed at, (a) low rotating speed (7,000 rpm) and (b) high rotating speed (11,500 rpm).
SEM images of cross section of sintered zirconia, (a) at bottom (7,000 rpm), (b) at middle (7,000 rpm), (c) at top (7,000 rpm), (d) at bottom (11,500 rpm), (e) at middle (11,500 rpm) and (f) at top(11,500 rpm).
Inhomogeneity of the sintered bodies (Figs. 10 and 12) was the result of the inhomogeneous powder pacing shown in green compact. Defects or inhomogeneity introduced in earlier process steps are difficult to recover and remove in later processing steps. Defects may have been introduced by inadequate preparation of the zirconia slurry; the primary grain size of zirconia (39 nm) was less than that of alumina (0.14 μm). SEM images showed agglomerated particles in the pre-sintered compact (Fig. 13). Further research is necessary to understand the source of the inhomogeneity.
SEM observation of agglomerated particles found in pre-sintered compact.
All-ceramic artificial teeth were produced by combing an HCP with a resin shell-mold made using a 3D printer. Slurries of alumina and zirconia fine powders were poured into tubes that contained the shell-molds. The slurries surrounded the inside and outside of the mold and the molds were rotated at between 7,000 and 11,500 rpm in a centrifuge (the HCP buried compaction method). Green compacts having the shape of teeth were obtained using this process. The compacts were subsequently air sintered to obtain the finished products. The results are as follows.
