1 Introduction

The article describes a research problem focusing on a number of issues related to the development of 3D printing technology using the robocasting and binder jetting methods, enabling the effective processing of oxide ceramics, e.g., Al2O3, and accompanying processes including the selection and preparation of the initial charge material, debinding, sintering, quality control, while ensuring their maximum utility on an industrial scale. The development of 3D printing technology using the robocasting and binder jetting methods will allow, above all, the production of innovative ceramic filters used in the production of advanced components, e.g., turbine blades, segments of steering apparatus, and elements of thermal barriers, installed in jet engines.

A typical schematic diagram of the investment casting process is shown in Fig. 1.

Fig. 1
figure 1

Schematic diagram of the investment casting process (Ref 1)

Full size image

In the case of this type of advanced components, the key is their high metallurgical purity, which can be obtained, among others, thanks to the use of ceramic filters during the casting process (Ref 1).

Generally, the starting material for their production is a polyurethane sponge, impregnated in subsequent stages with ceramic masses, which differ in relative volume, the chemical composition of the materials used, as well as the type of binder, surfactants, and antifoams. After applying the ceramics, the filters are burned off to remove the residual polyurethane sponge. It is widely known that the currently used filter technology results in the occurrence of large heterogeneities in terms of porosity and the lack of repeatability of thermal and mechanical properties. It has been shown that the currently used classic ceramic filters are characterized by high brittleness, lack of resistance to mechanical shocks, and impacts of a stream of liquid metal. All these factors pose a risk of introducing material from a damaged filter into the casting mold, which translates into contamination of the casting alloy, and thus the occurrence of casting defects such as nonmetallic inclusions (Ref 2, 3).

The task of ceramic filters is not only to filter liquid metal but also to absorb the impact energy of molten metal into the walls of the casting mold. It should be emphasized that the production process of foam filters does not ensure the repeatability of the geometry, and the shape of the pores is not controlled in any way (only their average size is determined). An opportunity to change this state of affairs, i.e., to obtain a repeatable structure of ceramic filters while ensuring full control of their geometric characteristics, is created by the use of 3D printing methods (RP—rapid prototyping) for their production (Ref 4, 5). One of them is the robocasting technology currently developed by CREATEC as well as the binder jetting method. It is assumed that the adopted concept of a radical change in the method of producing filters will allow to improve their operational properties, which will contribute to:

  • Reducing the risk of problems related to the flow of liquid metal by ensuring uniform filter geometry throughout its volume,

  • Improving the purity of the alloy as a result of increasing the mechanical strength of the filters and impact resistance, and limiting the possibility of metal flow next to the filter.

The limitation of the previously used 3D printing technologies of ceramic filters is, however, the thickness of the printed layer and the achievable dimensional accuracy. These criteria mean that currently printed 3D filters cannot be used for the production of the most advanced systems for the so-called micro-filtration, necessary in the production of monocrystalline parts.

It is assumed that the solution to obtain the required accuracy will be the use of the binder jetting technology (Ref 6). The method will allow the manufactured components to achieve dimensional tolerances close to those achieved in the production process by means of pressure injection of ceramics. Binder jetting will be an excellent tool for low-volume production and manufacturing of prototypes. Therefore, this article presents issues related to the technological process, from the selection of powder, through the development of the composition of the robocasting paste, to the production of prototype casting filters with a repeatable shape made using both technologies.

Obtaining an answer to the technological challenges defined above will allow for the identification of limitations not only with regard to the production of ceramic filters and other advanced casting components (including pins and cores), but also a number of other components manufactured using pressure injection technology.

Therefore, the research aim was to develop a concept and verify the possibilities of using two 3D printing technologies—robocasting and binder jetting—for the production of ceramic casting filters with improved technological and operational parameters compared to filters made by traditional methods.

2 Materials and Methods

2.1 Material Selection–Al2O3

Based on previously conducted tests and obtained results (Ref 7), from among various suppliers of Al2O3 powders available on the market, Al2O3—CL 370 powder, delivered by Almatis GmbH, was selected for the investigation. The specification of CL 370 is presented in Table 1.

Table 1 Specification of Al2O3-CL 370 powder
Full size table

2.2 Powder Preparation for 3D Printing Process

2.2.1 Robocasting—Paste Preparation

The influence of the mixer rotational speed, mixing time, temperature and humidity as well as air access on the paste form, including its homogeneity depending on the mixture composition, was tested. The approximate compositions of the mixtures were determined on the basis of the literature data. Seven variants of pastes were tested. Each time the components were weighed (the RADWAG PS1000.R2 precision balance), and then they were placed in a 500-ml container and mixed in a vacuum for two time variants, for 3 and 6 minutes. An MX4N vacuum mixer was used to mix the pastes.

The mixtures were made in a dozen or so variants with different chemical compositions. The influence of the mixer rotational speed, mixing time, temperature and humidity as well as air access on the paste form, including its homogeneity depending on the mixture composition, was investigated. The basic requirements for the paste intended for 3D printing were obtaining a homogeneous mixture, without agglomerates, deaerated, and with good consistency for squeezing. All parameters were checked organoleptically and in tests of squeezing through the printing nozzle.

2.2.2 Binder Jetting Technology

The basic and at the same time the most important element of printing in the binder jetting technology is the distribution of the powder/input material in the area of the working area. Depending on the type of material, the size and shape of the particles, and their moisture level, an individual approach is required to create dense layers with an evenly distributed powder (Ref 8,9,10). For example, powders with a particle size above 30 µm usually can be applied dry because they have good flow properties. For smaller powders, additional steps are often required to prepare the powder for the manufacturing process. The better the powder is distributed and the denser its packing, the better the properties of the final product. Usually, powders with a higher apparent density and a lower ratio of the bulk density to the apparent density have a better spread. The second step after powder selection is the selection of an effective binder, which is required to be low in viscosity, allowing it to easily detach from the nozzle during application. Additionally, the binder must be able to withstand the shear stresses acting during the process. The binder is sprayed onto individual layers of the powder and penetrates between the particles, binding them together. After printing, the binder is removed, therefore important parameters are also low chemical harmfulness, if possible it should be neutral to the environment, and ease of removal at the lowest possible temperature.

2.3 3D Printing

Figure 2 presents the filter geometry used to investigate the printing process with two technologies as well as the verification of suitability in casting tests. The filter diameter was 60 mm and the height depended on the number of layers applied, 10 layers were used. Then, using a 3D robocasting printer having a working area of 380 mm in diameter, a series of filters of eight pieces was produced. The smallest possible to be achieved with the use of the developed 3D printing technology, the thickness of the printed layer is 2.0 mm. It is strongly dependent on the diameter of the nozzle used. When implementing the technological process, one should take into account the phenomenon of paste expansion, as a result of which the diameter of a single stripe is greater than the diameter of the nozzle used. Therefore, the diameter of one strip was about 3 mm, and there were 11 of them on each layer.

Fig. 2
figure 2

CAD design of filter made of strips

Full size image

Using the aforementioned working area, during one setting of the 3D printing process, it was possible to perform 3D printing of five ceramic filters. The maximum printing speed, while ensuring acceptable technological characteristics of the paste (consistency, homogeneity, fluidity, continuity) was 30 g/min.

Filters are also made using binder jetting technology. Moreover, 10 mm cubes were printed, on which it was easier to observe the correctness of printing and control, for example, dimensional changes. The printing parameters are presented in Table 2.

Table 2 The most important parameters of the printing process using binder jetting technology
Full size table

2.4 Scanning Electron Microscopy

The printed and sintered metallographic sections of samples were observed by the scanning electron microscopy (SEM, model JSM-6460LV, JEOL) method. Microscopic observations were carried out in the modes of backscattered electrons (BES) and secondary electrons (SEI) at an accelerating voltage of 10 kV, using magnifications in the range of 1000x. The microstructures are shown in Fig. 3.

Fig. 3
figure 3

Microstructure of samples printed using the method: (a) binder jetting—on the left (b) robocasting—on the right

Full size image

2.5 Mechanical Test—Compression Strength Test

The compression strength tests were performed using a hydraulic press with a maximum load of 400 kN and an adjustable plunger speed ranging from 0.6 mm/min to 10 mm/min. The tests were carried out on samples formed using the binder jetting method, as well as samples formed using the robocasting method. The obtained results are summarized in Table 3.

Table 3 Comparison of results of compression strength test of samples after printing by binder jetting (BJ) and robocasting (RC) methods
Full size table

3 Results and Discussion

Based on experiments carried out related to the production of homogeneous pastes, for the robocasting technology, using various parameters, it was found that the homogeneity of the mixture was directly proportional to the rotational speed of the mixer and the mixing time. However, it was observed that the mixing time had a greater impact on the quality of the obtained pastes. Moreover, as expected, the longer the mixing time and mixer speed used, the more uniform the paste could be obtained. However, taking into account economic and production considerations related to the highest possible production efficiency and the lowest tool wear, it was found that for a sample with a volume of 500 ml, the acceptable level of paste mixing qualifying it for printing was achieved after 6 minutes of mixing using a mixer speed of 400 rpm. The composition of the tested mixtures is presented in Table 4.

Table 4 Composition of paste mixtures used for 3D printing filters using robocasting technology
Full size table

For each of the prepared mixtures for robocasting technology, its behavior during printing was verified. It was assessed, among others: its extrudability, printability, and print quality. A representative example of a test print for one of the analyzed mixture variants is presented in Fig. 4.

Fig. 4
figure 4

A representative example of a test print by robocasting technology

Full size image

Based on the assessment of the paste's behavior during printing and the organoleptic analysis of the prints, it was found that the best results were obtained for the mixture designated as no. 5, which was selected for further tests. In this way, a set of ready-made robocasting filters was obtained for further testing of the investment casting of critical aircraft engine parts.

On the other hand, the evaluation of the binder jetting printing process was carried out on cubic blocks and test filters with different strip thicknesses in order to verify their durability and technological capabilities of the process (Fig. 5). Regardless of the manufacturing technology used, the samples were first dried in a laboratory dryer according to the profile presented in Fig. 6, and then sintered in an HT16/18 Nabertherm oven in an air atmosphere according to the profile shown in Fig. 7.

Fig. 5
figure 5

Representative print of a cube and a filter with different strip thicknesses using binder jetting technology

Full size image
Fig. 6
figure 6

Filters drying temperature profile

Full size image
Fig. 7
figure 7

Filters sintering temperature profile

Full size image

The microstructure obtained by robocasting technology displays a material with numerous pores of irregular shapes, which suggests the presence of surface defects or inclusions resulting from production processes, such as degassing during sintering. The surface appears relatively homogeneous; however, the presence of pores may affect the mechanical properties of the material. On the other hand, the microstructure obtained by binder jetting technology presents a nonuniform rough surface resulting from the agglomeration of spherical particles, which vary significantly in size. The largest particles have a diameter of approximately 10 μm, with smaller particles adhering to their surface. Such a microstructure is typical for ceramic parts produced from sphere particles by classic cold compaction and sintering technology. The agglomeration and surface roughness of the particles influence on mechanical properties, such as compressive strength and material porosity.

Based on the strength tests carried out, it was found that samples formed by the binder jetting method were characterized by relatively low values of compressive stress, presented in Table 3. The highest compression stress, 15.4 MPa, was obtained for the sample described as a BJ_2. However, samples formed by the robocasting method exhibit significantly higher, by two orders of magnitude, compared to the binder jetting method values of compression stress. The highest values, 1718.2 MPa, were obtained for the sample described as a RC_2. The obtained results indicate that the robocasting method may be more effective for applications requiring high mechanical strength, whereas binder jetting may be suitable for less demanding environments.

On the basis of the conducted research, the differences between the technologies were defined and are presented in Table 5.

Table 5 Differences between robocasting and binder jetting technologies
Full size table

In summary, both in robocasting and binder jetting technologies, it is possible to produce filter structures of a repetitive shape. These results could be crucial for the correct design of the 3D printing technology of filters used in the precision casting technology of nickel superalloys.

The last stage of the research was technological research carried out in the foundry. Traditionally used filters allowed the mold to be poured in approximately 2.5 seconds. The produced Al2O3 strip filters met the main functional assumptions, i.e., durability of the structure and separation of contaminants, but extended the time of pouring the mold to over 3.5 seconds. This resulted in casting defects in the form of underfills. Therefore, further research will focus on optimizing the geometry of the filters to reduce the time of pouring the mold, but without reducing the filtration efficiency and thus guarantee the correctness of the manufactured details.

The use of fully repeatable filter production technology will also increase customer confidence that the product used for filtration is effective. Due to the homogeneous properties of the filters, it will be possible to treat the factors related to the alloy filtration as constant parameters, which is particularly important in the case of complex, multivariant analyses, often performed for the processes of investment casting of aircraft engine parts made of nickel superalloys. In addition, the use of the RP technique for the production of filters enables smooth adjustment of the metal flow rate, controlled by adding successive layers of the filter or changing the thickness of a single layer.

4 Conclusions

The article discusses the problem of using traditional ceramic filters in the investment casting of nickel superalloys and the related limitations. The authors presented the concept of using 3D printing methods—robocasting and binder jetting for the production of ceramic filters, which will offer higher quality than those produced in a classic way. The carried-out tests confirmed the possibility of producing filter structures of a repetitive shape using robocasting technology as well as binder jetting technology. These results could be crucial for the correct design of the 3D printing technology of filters used in the precision casting technology of nickel superalloys. Further research will focus on optimizing the geometry of the filters to reduce the time of pouring the mold, but without reducing the filtration efficiency and thus guarantee the correctness of the manufactured details.