Introduction

Additive manufacturing (AM) is a manufacturing process where objects are built up layer-by-layer. Formerly, AM was also referred to as rapid prototyping (RP) or solid freeform fabrication (SFF). In popular science, the term 3D printing is used as synonym for additive manufacturing. According to ASTM, additive manufacturing is a “process of joining material to make objects from 3D model data, usually layer upon layer” (Ref 1). Today, additive manufacturing of polymers is state of the art. In the field of metals, more and more materials can be processed as well. For processing ceramic materials, the technical application of AM technologies has been limited so far. However, ceramic materials have been studied in additive manufacturing processes ab initio with the development of the different AM technologies since about 25 years (Ref 2, 3). All established AM technologies have also been tested for ceramic materials (Ref 4-28). AM technologies can be classified according to the state of the material that is used (powder materials, liquid materials, and solid materials) (Ref 29, 30) and the deposition of the material (direct, indirect) (Ref 31). Direct means that the material is directly deposited only in the position giving the desired shape of the final object. These technologies like fused filament fabrication (FFF), direct inkjet printing (DIP), or thermoplastic 3D printing (T3DP) stand out due to the possibility to the AM of multi-material components (Ref 32, 33). The lithography-based ceramic manufacturing (LCM) is part of the indirect technologies (Ref 31).

Using AM as shaping technology increases the flexibility in regard to the design of components and enables innovative and maybe more effective structural designs, which could not be realized by conventional shaping. Until now, constructers design and develop components based on what is possible to produce and not based on the optimal desired functionality. However, with AM another strategy could be chosen in generating new functional- or structural-oriented components, e.g., catalyst supports, micro-reactors, static mixers (Ref 34), or heat exchangers. Looking at heat exchangers the process conditions, the fluids used, the temperature, and the pressure applied rule the used materials and their supports. Ceramics stand out based on an excellent reliability in regard to high temperatures, abrasion, and extreme chemical environments. Another important point is the ratio of heat exchanger surface area, the pressure drop within the structure, and the volume of the whole component. AM can be seen as door opener to produce complex geometries and channel systems in order to minimize the required volume and in order to increase the heat transfer surface. Simultaneously, the flow behavior inside the components can be optimized to avoid a high pressure drop.

We chose a simple example for a heat exchanger challenge: A hot fluid (e.g., hot aggressive chemical) runs through a pipe with an inner diameter of 10 mm, an outer diameter of 11.5 mm, and a length of 25 mm. This pipe has an outer surface of about 903 mm2 which can be used to transfer the heat to the surrounding fluid (e.g., air or water).

A common way to increase this heat transfer surface is by adding some rips to the outer surface. First, we chose eight rips with a constant cross section of 1 mm thickness and 9.25 mm height (Fig. 1, right). The resulting heat transfer surface is about 4452 mm2. But the heat flow within the single rips depends on the thermal conductivity of the used material and if this is too low the tops of the rips stay cold because the heat is transferred to the surrounding fluid in the lower areas. The surface at the tops of the rips is not used for the heat transfer, and not the whole available surface is used for the heat exchange.

Fig. 1
figure 1

CAD models of heat exchanger geometries and contact area to vat bottom during the building process—right sample 2, pipe with eight rips with constant cross section; middle: sample 3—pipe with eight rips with trapezoidal cross section; left sample 4—pipe with frustums instead of rips

Full size image

But it is possible to change the geometry of the rips to a trapezoidal cross section (Fig. 1, middle). Then more heat is transferred to the tops of the rips, and the whole available surface (about 4284 mm2 for eight rips with a height of 9.25 mm, a bottom thickness of 2 mm, and a top thickness of 1 mm) can be used for heat exchange. This surface is a little bit smaller than in the previous example, but more heat is transferred to the tops of the rips.

All these structures had a constant cross section for the whole component and can be manufactured with a technology like extrusion, slip casting, or injection molding (IM), which all have a very high productivity compared to other ceramic shaping technologies.

To increase the heat transfer surface further, we can separate the single rips into different frustums (Fig. 1, left). If each of these 96 frustums (8 × 12) has a height of 9.25 mm, a diameter of 2 mm at the bottom, and a diameter of 1 mm at the top, a heat exchange surface of about 4783 mm2 results. This surface is about 11.7% higher than the surface of the example with the trapezoidal rips. The production of this structure is more complicated. If extrusion is used, a machining step is needed afterward. For IM or slip casting, the geometry is very complex and therefore is likely to include some defects in the frustums in case not all channels in the used mold can be filled perfectly. Furthermore, the demolding of these structures is very complicated.

AM technologies make it possible to manufacture structures like this. But to utilize the opportunities of AM technologies, the geometry was designed more complex. Figure 2 shows a possible structure with thin bent frustums which are connected to each other by other added frustums. We calculated a resulting heat transfer surface of about 7441 mm2 which is more than eight times higher than the surface of the simple pipe and about 56% higher than the surface of the structure with the frustums.

Fig. 2
figure 2

Heat exchanger—sample 5—possible heat exchanger structure only producible with AM; CAD model

Full size image

Experimental

LCM Technology

We used the LCM technology (Ref 31, 35, 36) (LCM machine: CeraFab 7500, alumina suspension Lithalox 350, both of Lithoz-GmbH, Vienna, Austria) to manufacture all structures except the simple pipe.

The LCM technology differs from the well-known stereo-lithography by some features. A blue light source with a wavelength of 465 nm and a digital light processing (DLP) projector is used instead of UV light laser to cure a photosensitive suspension which mainly consists of monomers, photoinitiators, and homogenous dispersed ceramic particles. The upside-down setup of CeraFab 7500 is a big advantage, because the non-cured suspension runs back into the suspension reservoir and can be used for the production of the next layer. This means that only a small volume of suspension is necessary as well as that the cleaning process of the parts is easier because only a small amount of non-cured suspension remains in the manufactured part during the building process. The exposure of the first deposited layer to light occurs by backward located LEDs, inserted to the building platform. This cured layer is completely adhered to the building platform and guarantee fixation during the complete manufacturing time of the component. All subsequent layers are irradiated from the bottom below the building platform by DLP irradiation throughout the glass bottom of the rotating vat. We described the LCM process in a former paper in detail (Ref 34).

Because of the high resolution (635 dpi, pixel size 40 × 40 µm2) and the adjustable layer thickness (5-100 µm), a very good surface quality (compared to other AM technologies) and very thin structures (minimal wall thickness 100 µm) can be realized.

Necessary Design Changes

The pipes with the rips were manufactured with the same designs which were used to calculate the heat exchanger surfaces. But for the pipe with the frustums a little change in the design was necessary to avoid supporting structures between the frustums of one row. This was realized by changing the orientation of the frustums a little bit. The angle between the rotational axes of the inner pipe and the frustums was changed from 90° to 60° (Fig. 1, left).

For the last, very complex structure we added a support structure below the first frustums to support these during the printing process and to increase the connection between the whole structure and the building platform. These supporting structures were removed from the green components after the manufacturing process.

Last but not least, all edges were rounded to avoid crack formation in the ceramic components.

Building Process

For the building process, the parameter set given by Lithoz for 25 µm layer thickness and “middle” thick structures with 3.7 s exposure times were applied. The velocity for removing the cured layer from the tube bottom was decreased by 30%.

Cleaning, Debinding, and Sintering

For cleaning we used the cleaning fluid of Lithoz combined with compressed air. The non-cured suspension was removed, which remains on the structures because of the cohesiveness. For the thermal debinding, we followed the instructions given by Lithoz with different dwell times at 75, 115, 205, and 430 °C and a maximum temperature of 900 °C (Table 1). This temperature is necessary for first pre-sintering processes to increase the mechanical strength of the structures. The sintering happened at a temperature of 1600 °C with a dwell time of 2 h (Table 2).

Table 1 Used debinding regime
Full size table
Table 2 Used sintering regime
Full size table

Results

Figure 3, 4, 5, 6, and 7 show the different sintered alumina structures. Delamination defects in the rips between two layers were not visible after the manufacturing process. After debinding first little defects became visible and during the sintering process these defects increased further (Fig. 3 and 4; defects marked by arrows). The structure with the frustums as well as the pipe with the complex AM structure was sintered defect free and showed no delamination (Fig. 5, 6, and 7).

Fig. 3
figure 3

Heat exchanger—sample 2—pipe with eight rips with constant cross section; sintered alumina component with typical defects

Full size image
Fig. 4
figure 4

Heat exchanger—sample 3—pipe with eight trapezoidal rips; sintered alumina component with typical defects

Full size image
Fig. 5
figure 5

Heat exchanger—sample 4—pipe with frustums instead of rips; defect-free sintered alumina component

Full size image
Fig. 6
figure 6

Heat exchanger—sample 5—pipe with complex AM structure; defect-free sintered alumina component—horizontal view

Full size image
Fig. 7
figure 7

Heat exchanger—sample 5—pipe with complex AM structure; defect-free sintered alumina component—top view

Full size image

Discussion

Designing and Manufacturing Process

It is not possible to manufacture all geometries with AM technologies. For each AM technology, different limitations exist. Compared to powder bed-based processes which can use the powder bed as support, all cured areas of the manufactured layer have to be connected to the building platform/the last manufactured layer for the LCM technology or within the cured layer. Overhangs can be realized up to 60°, and outstanding structures are critical. But it depends on the size of the overhanging area and the distance between the connection points.

Figure 3 and 4 show typical defects of LCM components. The cracks are orientated parallel to the manufactured layers. The probability of generating these cracks increases with the thickness of the manufactured structures.

A possible explanation could be that for bigger structures the contact area between the vat bottom and the manufactured layer is bigger than for fine structures (Fig. 1). During the detaching process of the manufactured layer from the vat bottom, mechanical stress is induced into the structure and this stress increases with the size of the contact area because of the significant higher adherence on the vat bottom. This initiation of mechanical stress happens for every layer, and if this induced mechanical stress is too high, some non-visible delaminations are possible, which become visible after debinding and/or sintering.

The contact area between the vat bottom and the manufactured layer is significantly smaller for the structure with the frustums compared to the structure with the rips. The induced mechanical stress is lower, and no defects were initiated.

Possible solutions could be to decrease the exposure time to decrease the adherence between the structure and the tube bottom (this should be possible for these thick walls) or to slow down the detaching velocity.

Another challenge is the stable connection between the components and the building platform. The mechanical stress which is induced during the detaching process can result in tearing the components off the building platform. To avoid this defect, the contact area between the building platform and the first layer of the component and the support structure should be larger than the contact area between the cured area of the manufactured layer and the tube bottom but at most of the same size. Figure 8 shows three different configurations. The left one is a very critical one, because of the small contact area to the building platform and the big one to the vat bottom. By adding a support structure (middle), the connection to the building platform can be increased significantly. The example on the right shows a non-critical component.

Fig. 8
figure 8

Connection to building platform (blue) vs. connection to vat bottom (red)

Full size image

For the manufacturing process of the support structure, the very same material has to be used as for the component. Therefore, the removing of the support structure from components with fine structures is very critical. If possible supporting structures should be avoided by changing the orientation of the component in regard to the building platform or changing the component design to realize functional geometries which work as supporting structure as well (e.g., the rips of the component in Fig. 9 which acts as supporting structure and later as heat exchanger surface).

Fig. 9
figure 9

Heat exchanger—sample 7—alternative heat exchanger structure; defect-free sintered alumina component

Full size image

The LCM technology allows the AM of components with very fine structures and very good surface qualities. We have manufactured honeycomb structures with a wall thickness of 100 µm and a channel cross section of 0.9 × 0.9 mm2 as structure with the thinnest walls.

Holes with very small diameters (<200 µm) tend to clocking. The non-cured suspension remains in this area and is cured during the manufacturing of the next layers because of scattering light. This material cannot be removed during the cleaning process and is sintered like surrounding suspension material.

A critical point is the manufacturing of components which combine fine and thick structures in one layer. To cure very small areas and very big areas in one layer, different process parameters ought to be used. Small areas have to be cured with a high intensity and long exposure time, for big areas the used energy has to be lower to avoid overexposure, geometry deviation, and strong adherence to the vat bottom.

But this is not possible with the current device, and such geometries should be avoided.

Cleaning

The biggest challenge during the cleaning process is to remove the non-cured suspension completely. Most but not all can be removed with compressed air but also a cleaning fluid can be used. But if too much cleaning fluid is used or the component is exposed to the cleaning fluid for a prolonged time (>5 min), swelling occurs and defects are introduced to the components. New cleaning procedures are required for the future, to prevent swelling and defects introduction during cleaning. Another challenge is the handling of complex components to avoid any defects during the cleaning process.

Debinding and Sintering

The debinding regime given by Lithoz needs about 6-7 days for the alumina suspension depending on the realizable cooling rate. This time could be reduced by intensive investigation of the debinding behavior.

During the sintering process, deformations could occur for complex or thin structures. Possible supporting structures and changes in geometry have to be implemented during the designing process.

Costs

The production costs for LCM components depend on the number of components, which are produced. The costs, which have to be considered and have to be spread across the number of components which should be manufactured, are:

  • time for investigation of optimal machine parameters (curing parameter, detaching velocity) and supporting structure

  • time for generation of specific machine file (array and orientation of parts, supporting structure, process parameters)

  • machine costs

  • utilization of the building area (the time needed for the production of one layer does not depend on the area which has to be cured → the building platform should be fully loaded if possible)

  • height of parts, selected layer thickness (the time needed to manufacture one layer is about 40-60 s depending on the selected process parameters)

  • time and energy for debinding and sintering → utilization of the kiln

Additionally, the variable costs, nearly similar for each part, have to be considered:

  • time for cleaning

    • effort depends on part geometry—outer surfaces much easier to clean than inner channels, big channels easier than thin ones

  • costs for cleaning equipment

  • suspension costs

    • suspension needed for the manufacturing of the part(s)

    • loss of suspension depends on part geometry—much more non-cured suspension remains in thin channels and complex geometries than on outer smooth surfaces

If a high number of parts should be processed, the development of handling aids and cleaning strategies could be reasonable. With these additional efforts, especially the time for cleaning and the number of defects can be reduced significantly.

For mass production the production costs of the LCM technology will always be higher than for conventional shaping technologies. Fast production of low quantities is suitable. But the future of this technology will be the production of components with very complex geometries.

Conclusion

AM technologies allow the realization of components with a complex design which cannot be produced with any other technology. For the AM of ceramic components, the LCM technology stands out because of the very good component properties, the high sinter density, and the high resolution (smallest wall thickness about 100 µm).

To use the new degree of flexibility in design, a mind-changing process has to be initiated and simulation tools have to be used to design the components to function (flow, heat transfer, mechanical strength, etc.). But the opportunities of the known CAD tools are limited and further developments are needed to realize a designing tool which allows easy designing and changing of complex structures according to the simulation results. At the end a modification of the CAD model is necessary to optimize the geometries for the AM process. Preferably this necessary step is implemented into software tools which consider the specific restrictions of the chosen AM technology.

If these challenges could be solved, the AM technologies are one key for functionalization and miniaturization of ceramic components. Figure 9 and 10 show a possible heat exchanger structure manufactured by the LCM technology and an alumina suspension. A heat transfer surface of more than 3500 mm2 could be realized in a structure with an outer diameter of 26 mm, a height of 13 mm, and an inner diameter of the pipe of 2.2 mm. The surface can be increased further by adding rips or other structures to the top of the structure. The rips at the bottom were used as remaining support structure which connects the component to the building platform during the building process and which works as heat exchanger surface after sintering.

Fig. 10
figure 10

Heat exchanger—sample 7—alternative heat exchanger structure; defect-free sintered alumina component; top view

Full size image

For AM of ceramic components using the LCM technology, the following statements can be made:

  • during the design process of the component, the following things should be addressed to avoid any defects:

    • minimal wall thickness is about 0.1 mm

    • the minimal diameter of a hole depends on the aspect ratio, but holes with a diameter of 0.2 mm can be realized

    • thinner structures have small contact areas to the vat bottom and debinding can happen faster

    • significant differences in wall thicknesses should be avoided in one layer because of the homogenous exposure of all selected areas

    • the connection area to the vat bottom should be the same or smaller than the connection area to the building platform

    • check if functional structures can be used as supporting structures to avoid the removing step

  • during the manufacturing process:

    • for thin structures more exposure time and intensity are needed than for thicker structures

    • the velocity of the detaching process has to decrease with increasing size of the cured areas

  • during the cleaning process:

    • use compressed air for cleaning and reduce the usage of solvents (Ref 37)

  • debinding and sintering process:

    • check the real temperature distribution of the used kiln and modify the used temperature regime if necessary

    • the needed timeframe for debinding and sintering can be reduced by thermal analytics and adjustment of the temperature regimes

  • costs:

    • the costs per part decrease with the number of parts

    • compared to other technologies like ceramic injection molding:

      • the costs for a small number of parts (<50) are significant lower

      • the costs for a large number of parts are significant higher

      • the breakeven point depends on the component size, the used ceramic material, and the material used for the mold