Binder Jetting Additive Manufacturing: The Effect of Feed Region Density on Resultant Densities

Many applications such as tooling and biomedical implants can benefit from binder jetting (BJ) additive manufacturing. A wide range of researchers have reported their studies on BJ [1–13]. Many published studies are on how final density (usually the sintered density) in the BJ process is affected [4–13].

A common method for altering the final density is to alter the intermediate densities (powder bed density and green density) [4,5]. The work by Li et al. indicated that the powder bed density approximately equaled feed region density [13]. According to this result, a natural conjecture is that increasing the feed region density (e.g., by compacting the powder in the feed region) might increase the powder bed density (and consequently green density and final density). In fact, some 3D printer manufacturers suggest that operators should compact the powder in the feed region to improve the final density. However, the literature does not have experimental data to substantiate the conjecture. It is not clear how powder spreading affects this conjecture because the spreader (roller or blade) disrupts the original particle packing and rearranges the particles when it moves the powder from the feed region to the build region (in which the powder bed is located). The previous work by Li et al. [13] was conducted without any compaction; therefore, it is still unknown how increasing the feed region density affects the final density and the intermediate densities (powder bed density and green density), especially considering the particle rearrangement during powder spreading.

This study aims to experimentally determine whether compacting the powder in the feed region can result in a corresponding increase in the powder bed density, green density, and final density (sintered density). The experiments were carried out on a commercially available BJ 3D printer. Statistical testing was conducted on the experimental data.

This study was conducted on alumina powder with a nominal particle size of 20 µm provided by Inframat. A laser scattering machine (HORIBA LA-960) was used to measure the particle size distribution of the alumina powder, which was repeated thrice. A scanning electron microscope (SEM; JEOL JSM-7500F) was used to characterize the morphology of the alumina powder.

A commercially available BJ 3D printer (ComeTrue T10, Microjet Technology) was used for conducting experiments. To reduce the amount of powder needed for the experiments, two sets of custom devices, which were created in the previous work [13], were installed on the 3D printer. Figure 1 depicts how the two custom inserts were placed above the original feed region platform and the original build region platform of the 3D printer and aligned with the top of the feed region and the build region. The two custom pistons with a size of 25 × 25 mm² for the feed region and build region were fixed to the original platforms, which allowed the custom pistons to follow the downward and upward movements of the original platforms during the printing process.

The specimens were debound and sintered with the following process: heating from room temperature to 500 °C at 3.2 °C/min, heating from 500 °C to 1150 °C at 2.4 °C/min, heating from 1150 °C to 1700 °C at 0.8 °C/min, dwelling at 1700 °C for 120 min, and finally cooling in the furnace to room temperature via natural convection. They were then removed from the furnace for investigation.

The resulting density data were then put through a statistical analysis to determine whether there are significant effects of feed region density on powder bed density, green density, and sintered density.

The sintered specimens were mechanically fractured to reveal the interior surface for microscopy. Specimens were imaged with the tescon vega ii lsu sem.

Figure 2 presents the three trials for the particle size distribution measurement of the alumina powder with the laser scattering machine. The mean size is 26.4 µm with a median of 24.8 µm, so the particles are slightly larger than the nominal size of 20 µm on average.

The morphology of the powder is shown in Fig. 3. The particles are largely spherical and do not appear to agglomerate. The lack of agglomerations indicates higher likelihood for good spreadability.

Figure 4 shows the measured powder bed density, green density, and sintered density as a function of feed region density. An increasing trend is captured and depicted for each phase of the BJ process. A higher feed region density correlates with a higher powder bed density, green density, and sintered density.

To determine whether feed region density has a statistically significant effect on powder bed density, green density, and sintered density, three analysis of variance tests were conducted. The sample size was three for analyzing the powder bed density at each feed region density. For the green density and sintered density analysis, the sample size was eight at the feed region density of 44% and was seven at the feed region densities of 55% and 65% because of one of the eight specimens was damaged in these two cases. The p-values are 0.0015, 0.0009, and 0.0002 for powder bed density, green density, and sintered density, respectively. The small p-values are far lower than the commonly accepted significance level of 0.05 for the three tests, verifying that the feed region density has a significant effect on each of the three investigated phases of the BJ process.

The mean powder bed densities for the three investigated feed region densities, 44%, 55%, and 65%, were compared with the data provided in Table 2. The p-values for the comparison of 44% and 55%, 55% and 65%, and 44% and 65% are 0.038, 0.034, and 0.001, respectively. The p-values being lower than 0.05 validates that mean powder bed densities are statistically different from one another with greater than 95% confidence.

The mean green densities for the three investigated feed region densities, 44%, 55%, and 65%, were compared with the data provided in Table 3. The p-values for the comparison of 44% and 55%, 55% and 65%, and 44% and 65% are 0.015, 0.012, and 0.003, respectively. The p-values being lower than 0.05 validates that the mean green densities are statistically different from one another with greater than 95% confidence.

The mean sintered densities for the three investigated feed region densities, 44%, 55%, and 65%, were compared with the data provided in Table 4. The p-values for the comparison of 44% and 55%, 55% and 65%, and 44% and 65% are 0.048, 0.017, and 0.0002, respectively. The p-values being lower than 0.05 validates that the mean sintered densities are statistically different from one another with greater than 95% confidence.

The powder bed density was increased by 5.7%, the green density was increased by 8.5%, and the sintered density was increased by 4.5% as a result of increasing the feed region density from 44% to 65%. The likely reason for this increase in these densities is the increase in the effective dosing ratio. By compacting the feed region, more powder is made available for the spreading of each layer without increasing the layer height although the nominal dosing ratio is the same. However, the future work is required to confirm or reject this likely reason.

Figure 5 shows the fracture surfaces of the sintered specimens at three different magnifications. All the sintered specimens show a low amount of necking and clear granular structures. This is mainly because of the large particle size and thus the low sinterability of the powder. The different densities at various feed region densities are difficult to identify visually, but the third row of Fig. 5 (at the highest density in this study) shows slightly denser packing of the sintered particles.

This experimental investigation into the effects of feed region density on the density of the powder beds, green specimens, and sintered specimens showed that they were significant. A higher density in the feed region resulted in a higher density for the specimens throughout the BJ process. The statistical testing verified that the differences in the mean powder bed density, green density, and sintered density were significant with greater than 95% confidence. By increasing the relative feed region densities from 44% to 65%, the powder bed density, green density, and sintered density was increased by 5.7%, 8.5%, and 4.5%, respectively. Compacting the powder in the feed region before printing can alter the sintered density of the product produced by BJ. This can be beneficial for various applications where density plays an important role, such as tooling and biomedical implants.

• National Science Foundation (Grant No. 2047908).

There are no conflicts of interest.

The authors attest that all data for this study are included in the paper.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.