1 Introduction

There has been an increase in the global market for miniaturised products across a range of industries such as the pharmaceutical, automotive and energy sectors. Such miniaturised products often necessitate the use of complex meso, micro- and nano-scale structures to satisfy their specific functional and technical requirements. This in turn requires enhanced component fabrication technologies and the development of new process chains which can manufacture such components in a reliable and cost-effective manner [1, 2].

Microinjection moulding (MIM) of thermoplastic polymers is one of the most reliable and cost-effective large-scale replication technologies available, with proven capacity in the manufacture of microfluidic devices and MEMS systems [3]. The advantages of the process include shape complexity, replication fidelity, net-shape forming, availability of a wide range of thermoplastic materials, and is suitable for large scale production [4,5,6]. Despite these advantages, a major limitation of the MIM, is its inability to produce complex 3D structures [3] and internal geometries due to demoulding difficulties. This limitation constrains product designs and necessitates the need for post-processing assembly activities such as a mechanical fastening, adhesive or a welding to obtain the desired final product [5, 7]. The current approach to produce the device involves a number of steps as shown in Fig. 1, namely splitting the design of the device into individual layers which could be separately injection moulded. The demoulded separate layers are then aligned and assembled into a microfluidic device using an appropriate adhesive. Such an approach to manufacture these parts can be time-consuming, costly, and labour intensive [3, 6, 8].

Fig. 1
figure 1

Typical manufacturing process of a microfluidic devices

Full size image

In order to avoid the abovementioned constraints, this paper is proposing a cost-effective approach to produce polymer meso- and micro-scale components with complex hollow internal geometries by combining 3D printing (3DP) and a variant of the MIM process into a process chain (See Fig. 2). This method will enable the fabrication of such components in a single replication step and eliminates the need of any post-fabrication assembly activities. A literature search has revealed that as yet, this technique has not been attempted for MIM of small-scale components and cavity features.

Fig. 2
figure 2

Proposed manufacturing process chain for plastic micro-components with internal hollow functional features

Full size image

As shown in Fig. 2, in the first step, the 3DP process will be used to fabricate lost-cores having the shape of the component’s hollow internal geometries. 3DP is deemed to be a promising candidate technology to fabricate the lost-cores because the process allows the fabrication of complex structures with precisely controlled geometry [9].

The lost-core technique has been widely used in the macro scale in numerous industries to achieve hollow internal complex structures without the need for any post-processing steps [10, 11].

3DP has been previously used to fabricate water soluble salt lost-cores in the macro scale for the manufacturing of hollow aluminium, magnesium alloy and composite castings [12, 13].

However, no studies have yet been carried out to investigate the use 3DP soluble plastic/polymeric lost-cores for over-moulding using polymer MIM process.

Two alternative 3DP technologies, namely fused filament fabrication (FFF) and digital light processing (DLP), can be considered as potential candidates for the first step in the proposed process chain. Both technologies have good printing capabilities and are compatible with commercially available soluble polymer materials from third-party sources. Furthermore, DLP printing offers distinct advantages over FFF due to its higher printing resolution [14, 15] as well as better surface quality and details achieved in the micro scale [16].

In the subsequent step, these lost-cores will be employed within the MIM process which will involve the over-moulding of another polymer around the 3DP fabricated lost-core. After demoulding of the polymer component, the lost-core is removed by dissolving it in its solving medium, resulting in the polymer components with the respective internal hollow geometries.

As discussed by Amin et al., 3DP has been used successfully to directly manufacture the entire microfluidic device [17]. However, the use of support structures, particularly in DLP could negatively impact the print quality of the internal hollow cavities of the microfluidic device and at the same time are also difficult to remove satisfactorily. Thus, 3DP the microfluidic devices with precise small hollow structure is extremely difficult when compared to forming the core. Furthermore, the proposed process chain has a number of advantages. Firstly, in the case of small batch production, a faster throughput can be achieved since only the core is 3D printed, and secondly, in comparison to 3DP, injection moulding provides a broader selection of thermoplastic materials for the over-moulding of the final part, including biocompatible polymers.

Therefore, this paper presents an experimental study on the feasibility of the proposed process chain to identify the constraints and required conditions for the successful fabrication of components with meso–micro scale internal features using the lost-core technique.

2 Materials and methods

2.1 Materials

2.1.1 3DP soluble polymer materials

The selected lost-core materials have to withstand the high injection pressure and temperature involved in the injection phase of MIM. Experimental and simulation studies on the use of fusible cores in IM have revealed that the surface temperature of the core does not rise above 130 °C during the process, even though the polymer melt temperature attains 290 °C [18]. Therefore, based on these recommendations, it is evident that the glass transition temperature (Tg) of a candidate soluble polymer material for the cores should be higher than 130 °C, since a higher Tg reduces the likelihood that the core loses its rigidity and shape [19, 20], e.g. in such an over-moulding process.

Another study by Gheisari et al. on the use of polymers for rapid IM tooling inserts also highlighted the importance of having a material with a sufficiently high Tg in order to prevent insert failure because of a decrease in tool strength with an increased IM temperature [21]. Furthermore, as highlighted by Dempsey et al., the heat deflection temperature (HDT) is an important material property for polymer-based IM rapid tooling. In particular, IM inserts made from polymers with a HDT below 95 °C failed during the injection phase of the IM process [22]. It is worth noting that, for many polymers, HDT is related to Tg since the polymer molecular chains start to become mobile, and thus, the material is easily deformable.

Apart from the temperature aspect discussed above, the solvents for the chosen core materials should have a low viscosity to allow for easy dissolution of the lost-core material within the small intricate cavities of the over-moulded component. Within this context, the soluble polymer materials chosen for this study include a commercially available resin-based material and two filament-based materials. Specifically, these are the IM-HDT-WS with a Tg of 190 °C from 3DResyns [23], AquaSys180 with a Tg of 171 °C from Infinite Material Solutions [24] and Xioneer VXL130 with a Tg of 130 °C from Belland Technology [25], respectively. All the three selected materials are compatible with off the shelf 3D printers and utilise water or a water-based solvent for the dissolution of the material after printing. Other mechanical properties of the abovementioned materials can be found in Table 1 below.

Table 1 Mechanical properties of the 3DP materials [23,24,25]
Full size table

2.1.2 Over-moulding material

The thermoplastic material selected for the over-moulding of the 3D printed cores during the injection moulding process was polymethyl methacrylate (PMMA) Plexiglass grade 8N from Röhm GmbH (Darmstadt, Germany) [26]. This material was chosen for its good flow, high mechanical strength, and excellent transparency due to its amorphous structure [27]. Furthermore, due to its good mechanical properties, it is a good candidate material for the fabrication of microfluidic devices [28, 29]. The mechanical and optical properties of the Plexiglas 8N are shown in Table 2.

Table 2 Mechanical and optical properties of Plexiglas 8N [26, 29]
Full size table

2.2 Methods

2.2.1 3D printing processes

DLP printing involves exposing the liquid photopolymer resin which is found in a vat in the printer to a UV light source to produce the parts. The light source is a specially developed digital light projector screen which flashes an image of an entire layer all at once. The part is generated on a layer-by-layer basis onto a build platform which moves independently from the resin bath. The DLP printer selected for this study to print the soluble lost-cores is the Asiga Max X27 (Sydney, Australia) printer for the resin-based material.

In contrast to the DLP which is a vat polymerisation printing method, FFF printing is a material extrusion (MEX) 3D printing method. In FFF, a thermoplastic material in form of a filament is melted in the printing head and extruded through a heated nozzle to produce the part layer by layer. The print head is moved in the XY plane in order to build the part onto the build platform of the printer. Once a layer is completed, the build platform is lowered in the z direction in order to start the next layer. The process continues until the fabrication of the part is completed [16, 30, 31]. In this case, the selected FFF printer is the Ultimaker S5 (Zaltbommel, Netherlands).

These printers were chosen for their printing resolution and for their compatibility with the selected third-party soluble polymer materials. Specifically, the Asiga Max X27 printer has a xyz resolution of 27 × 27 × 1 µm, [32], whilst the Ultimaker S5 printer has a xyz printing resolution of 6.9 × 6.9 × 2.5 µm [33].

Based on the range of 3DP parameters recommended by the respective materials suppliers, some preliminary trials were conducted to identify an appropriate combination of printing parameter settings to produce cores of the required quality. The selected 3DP parameter settings for the soluble polymer materials are shown in Tables 3 and 4. Furthermore, the resin material (IM-HDT-WS) necessitated post-processing steps which included, a UV post-cure for 20-min and a thermal post-cure of one hour at a temperature of 90 °C as recommended by the material manufacturer.

Table 3 DLP printing parameters for IM-HDT-WS resin [23]
Full size table
Table 4 FFF printing parameters for Xioneer VXL130 and AquaSys180 filaments [24, 25]
Full size table

2.2.2 Test artefact core design

The test artefact core design chosen for this study was a Y-channel mixing device as it is a commonly used microfluidic mixing device in the medical and pharmaceutical industries. The device is defined by having two inlets at a specified bifurcation angle (30 to 180°), and a single outlet from which the mixed fluid is outputted [34, 35]. The test artefact designed included Y-channel diameters of 1.6 mm and 0.8 mm with a bifurcation angle of 60°, as can be seen in Fig. 3. The design also included rectangular supports at each end of the Y-channel branches to support and locate the test core artefact when placed inside the mould cavity. This ensured the correct alignment of the core and allowed for the over-moulding material to flow freely around the artefact without any obstructions during the IM process step.

Fig. 3
figure 3

Y-channel dimensions with 0.8 mm and 1.6 mm diameter

Full size image

For ease of reference throughout this paper, the Y-channels will be referred to as Y1.6 and Y0.8 for the 1.6 mm and 0.8 mm diameter Y-channels, respectively.

2.2.3 Mould fabrication

A balanced four-cavity mould design was selected for the study, from which a central runner would simultaneously feed the polymeric material to all the cavities. This type of design ensures that each core in the cavity experiences the same moulding conditions during the injection process. The mould cavities were cut on an unalloyed tool steel insert from Meusburger (1.1730) using a Haas MiniMill CNC milling machine. The machined mould with the corresponding cavities can be seen below in Fig. 4.

Fig. 4
figure 4

Machined mould’s runner system and cavities

Full size image

2.2.4 Micro-injection moulding

Using the abovementioned mould, the over-moulding of the 3DP cores was conducted using a Boy GmbH & Co. KG (Neustadt-Fernthal, Germany) 22E IM machine with a 22 mm diameter screw and maximum clamping force of 220 kN. A temperature controller from Singer (USA) (STL 150/1-4-20-K6) was used to control the temperature of the mould, and a Helios GmbH (Germany) WINneo hopper-type drier was utilised to dry the PMMA material prior to moulding. The polymer was pre-dried for three hours at 90 °C as recommended by the manufacturers to ensure that all moisture present in the material is removed prior to IM process. The effective filling performance of MIM relies on correctly selecting the injection speed, melt temperature, mould temperature, holding pressure, and holding time [3, 36, 37].

Since the main objective of this research was to demonstrate the feasibility of the proposed process chain, an optimisation of the MIM process was not conducted. Therefore, the final processing parameters were selected based on the results reported in the above-mentioned studies, processing parameters recommended in the Plexiglas 8N data sheet, and some preliminary IM trials conducted in this study.

These preliminary IM trials were conducted to identify an appropriate combination of parameter settings to produce PMMA replicas with hollow geometries of the required quality whilst at the same time endeavouring to minimise the thermal and structural loads on the core. Based on these preliminary moulding trials, the final IM process settings which were eventually used in this feasibility study to over-mould the 3D printed soluble cores are shown in Table 5.

Table 5 Over-moulding processing parameters for PMMA (Plexiglas 8N) [26]
Full size table

2.2.5 Dissolution tests

The dissolution rates of the three chosen soluble polymer materials were studied by carrying out dissolution trials using printed cuboid shaped coupons with dimensions of 10 × 10 × 10 mm. An ultrasonic bath BIOSONIC Ultrasonic Cleaner (US-10SD) with temperature control operating at 40 kHz was used to agitate the solvent. Each of the cuboid samples were placed separately in a 200 ml beaker and then placed in the ultrasonic bath. A total of three samples for each of the soluble polymer materials was considered. The initial weight of each cube was measured using a precision balance (Precisa 404A, Mitutoyo) and the weight change of each sample was measured at 10-min intervals.

The solvent used for the IM-HDT-WS and the AquaSys180 was tap water whilst a proprietary powder, Xioneer VXL EX, mixed with tap water with a concentration of 2:1 to the weight of the material being dissolved was used for the Xioneer VXL130. The solvent temperature was set at 80 °C as recommended by the material manufacturers.

Furthermore, the same dissolution conditions described above were used for the dissolution of the over-moulded lost-core after the MIM process.

2.2.6 Characterisation

Various tests were performed on the raw materials, the 3D printed core test artefacts, and the over-moulded parts to evaluate the material properties, the geometrical accuracy of the 3DP technology, and the over-moulding process.

2.2.6.1 Microscopy

The Y-channel diameter and the bifurcation angle were analysed using a Bruker Alicona (Austria) InfinteFocus G5 Plus microscope to evaluate the geometries of the 3D printed part prior to over-moulding. A total of three printed parts of each core candidate material were measured against the test artefact’s design drawing.

After the over-moulding process, a total of three PMMA parts (with embedded cores) were mounted in 30 mm diameter plugs of thermosetting polymer for convenience in handling and ease of inspection. The plugs were then ground (coarse and fine grinding) and polished to obtain a clearly defined and smooth surface of the middle section of the PMMA parts. This allowed for satisfactory microscopic examination of the middle section of the over-moulded lost-core embedded within the PMMA material.

After dissolution of the cores, the replicated geometries of the soluble lost-core onto the over-moulded PMMA were also measured using the same microscope and procedure.

2.2.6.2 Differential scanning calorimetry (DSC)

A DSC analysis was performed using a Mettler Toledo (USA) DSC3 + [38] to determine the glass transition temperature (Tg) of the materials being investigated. The DSC used 99.99% Ni gas for purging at a flow rate 60 ml/min and a heating and cooling rate of 10 K/min. Each specimen was cycled through their heat-cool-heat cycle profiles and the temperature range for the three materials investigated was between 25 and 300 °C. Two specimens were studied for each of the FFF materials (Xioneer VXL130 and AquaSys180) which were kept in their raw state (filament) whilst two specimens for the DLP material (IM-HDT-WS) were printed and post-cured.

2.2.6.3 Dynamic mechanical analysis (DMA):

DMA analysis was performed using a 3-point bending on FFF and DLP 3DP specimens using a Mettler Toledo (USA) DMA1 [39] to measure the thermomechanical properties of the chosen materials. A total of three 50 × 5x2 mm specimens were considered for each material (FFF and DLP). The Xioneer VXL130 specimens were heated from 25 to 145 °C, the AquaSys180 were heated from 25 to 190 °C and the IM-HDT-WS specimens were heated from 25 to 210 °C at a frequency of 1 Hz and displacement of 10 µm.

3 Results and discussion

This section discusses the results obtained from the tests conducted on the 3DP materials, and the proposed process chain steps to obtain a polymer part with hollow internal geometries. The values of the channel diameters and bifurcation angles of the 3DP and replicated features, are used to quantitatively compare the results after each step of the proposed process chain.

3.1 Differential scanning calorimetry and dynamic mechanical analysis

Table 6 provides the data obtained from the DSC experiments for the 3D printed IM-HDT-WS as well as AquaSys180 and Xioneer VXL130 filaments. The values obtained from the DSC test conducted on the filament materials seem to be within the Tg range provided by Xioneer VXL130 and AquaSys 180 manufacturers, respectively. However, the Tg value obtained for the resin-based material (IM-HDT-WS) is 10.5% lower when compared to the material’s data sheet value specified earlier. This could result from a not optimum post-curing process of the printed lost-cores. The lower Tg value of the IM-HDT-WS material could negatively impact the performance of the core as it can lose its rigidity and shape during the injection phase of the over-moulding process.

Table 6 Glass transition temperature of soluble lost-core materials
Full size table

Figure 5 shows the DMA thermograph of each material. The complex modulus values displayed in the curves are the average values of three specimens tested. From the results of these DMA experiments, it can be observed that as the temperature increases, a decrease in the complex modulus of the material is observed. At a temperature of 75 °C, the AquaSys180 shows a sharp decrease in the complex modulus where the material loses its strength. On the other hand, the Xioneer VXL130 and IM-HDT-WS materials show a more gradual decrease in their complex modulus, where the IM-HDT-WS loses its strength at a temperature of 95 °C whilst the Xioneer VXL130 loses its strength at a temperature of 130 °C. The loss of strength in the IM-HDT-WS may be due to the post-curing process not being fully optimised. Table 7 summarises the results obtained from the DMA analysis, namely the variation of the complex modulus for specific temperature intervals.

Fig. 5
figure 5

DMA curve for FFF and DLP materials

Full size image
Table 7 DMA results of the soluble 3DP Materials
Full size table

Based on the measured values of the complex modulus of the chosen materials shown in Table 7, the best candidate core material is the Xioneer VXL130 as it retains the highest value of the complex modulus across the 50 °C to the 110 °C range.

3.2 3.2. Geometrical and dimensional tests

3.2.1 3D printed cores

All cores were satisfactorily printed using the respective printers according to the test artefact design. Figure 6 shows the Xioneer VXL130 core 0.8 mm channel diameter printed using the Ultimaker 5 FFF printer.

Fig. 6
figure 6

FFF core using Xioneer VXL130 (build direction is upwards towards the reader)

Full size image

Table 8 shows the average values of the diameter and bifurcation angles for the Y1.6 and Y0.8 channel cores for the Aquasys180, Xioneer VXL 130, and IM-HDT-WS 3DP cores.

Table 8 Y-Channel dimensions of printed FFF and DLP cores
Full size table

From the obtained measurements, it can be noted that there are some differences in the diameters of the channels when compared to the target values. The diameter differences observed in the Y1.6 channel are + 5.63% for the AquaSys180, + 1.88% for the Xioneer VXL130 and − 0.63% for the IM-HDT-WS. The differences observed in the Y0.8 channel are + 21.3% for the AquaSys180, + 5.0% for the Xioneer VXL130 and + 2.50% for the IM-HDT-WS.

Differences may also be noted in the bifurcation angle of the 3DP parts against the target values. The differences observed in the Y1.6 channel are -5% for the AquaSys180, + 5% for the Xioneer VXL130 and + 5% for the IM-HDT-WS. The differences observed in the Y0.8 channel are 0% for the AquaSys180, + 1.67% for the Xioneer VXL130 and + 1.67% for the IM-HDT-WS.

These relatively small differences in dimensions and angles can be attributed due to the printing resolution of the printers, most particularly the limited resolution of the FFF printer when compared with the DLP printer. Furthermore, this could be due to the geometrical compensation taken into consideration when designing the parts before being printed [40,41,42,43] and also potential measurement errors as the core channel edges were difficult to precisely detect.

Comparing the average results for the diameters and bifurcation angles for each soluble material and printer combination with the core design target values respectively, it can be observed that overall, they concur quite closely for both channel sizes. Furthermore, as expected the IM-HDT-WS material and DLP printer combination have produced the 3D cores with the best quality.

The abovementioned deviations from the nominal dimensions can be reduced by optimising the 3D printing process and increasing its accuracy. As previously reported, this is only a feasibility study and consequently the 3D printed cores for all three soluble materials were considered adequate to proceed with the analysis of the follow-up step in the process chain.

Upon visual inspection of the surface texture of the Y-channels, a noticeable difference can be seen between the cores produced by the FFF technology (Fig. 7a, b) and the DLP technology (Fig. 7c). The DLP technology achieves a better surface finish due to the higher achievable resolution and the method of printing utilised. This difference in surface texture may also be attributed to the difference in layer thickness between the FFF technology and the DLP technology [14, 44, 45].

Fig. 7
figure 7

Surface roughness comparison between a Xioneer VXL130 Y0.8, b AquaSys180 Y0.8 and c IM-HDT-WS Y0.8

Full size image

3.2.2 Over-moulded 3D printed cores

Figure 8 shows a 3DP lost-core encapsulated in PMMA after the MIM over-moulding process. Table 9 shows the average values for the diameter and bifurcation angels for the Y1.6 and Y0.8 channel cores for the Aquasys180, Xioneer VXL130, and IM-HDT-WS over-moulded cores.

Fig. 8
figure 8

Over-moulded lost-core—Xioneer VXL Y1.6 channel

Full size image
Table 9 Y-Channel dimensions of over-moulded parts using FFF and DLP cores
Full size table

When comparing the Y-channel dimensions (average diameter and bifurcation angle) prior to over-moulding as seen in Table 5, and the dimensions seen in Table 4 above, one can note a change, most notably in the bifurcation angle of the Y-channel.

The observed diameter differences of the Y1.6 channel are -7.64% for the AquaSys180, − 0.62% for the Xioneer VXL130, and 0% for the IM-HDT-WS. The differences observed in the Y0.8 channel are − 12.37% for the AquaSys180, − 2.38% for the Xioneer VXL130, and + 4.88% for the IM-HDT-WS.

Differences may also be noted in the bifurcation angle after over-moulding. The differences observed in the Y1.6 channel are + 12.28% for the AquaSys180, + 4.76% for the Xioneer VXL130, and − 3.17% for the IM-HDT-WS. The differences observed in the Y0.8 channel are + 13.33% for the AquaSys180, + 21.31% for the Xioneer VXL130, and + 11.48% for the IM-HDT-WS.

As expected, the greatest deflection of the lost-core is observed in the Y0.8 channel which is the smallest core diameter being tested. These large deflections may be attributed to the high temperatures and pressures experienced on the lost-core material during the MIM process [40] which were not optimised and due the location of the gate being located at the side of the core. One needs to also take into consideration potential measurement errors due to the Y-channel diameter edges not being well defined. Further research work is therefore required in order to optimise the MIM process, by taking into consideration the results and recommendations in literature [46,47,48,49,50] and thus improve the quality of the replicated internal cavities of the parts.

3.3 Dissolubility tests

3.3.1 Preliminary dissolution tests

The AquaSys180 cuboid coupons developed a gel-like residual layer which formed on the surface of the sample, as seen in Fig. 9. The Xioneer VXL130 and the IM-HDT-WS cuboid samples completely dissolved in their respective solvents without the formation of any gel-like residual layer. The results of the preliminary dissolution tests on the coupons can be seen in Table 10.

Fig. 9
figure 9

AquaSys180 coupon comparison, as printed (left), gel-like residual layer during dissolution (right)

Full size image
Table 10 Preliminary dissolution test results
Full size table

The Xioneer VXL130 had the fastest dissolution rate when compared with the other materials considered, whilst the Aquasys180 took the longest to dissolve. The prolonged dissolution time of the AquaSys180 could be attributed to the gel-like residual layer which could have negatively affected the dissolubility rate as it impeded direct contact of the solvent with the polymer material [51]. Furthermore, given the proposed application as a lost core, the gel-like residual layer may prove difficult to clean completely especially when the over-moulded parts have small complex internal geometries. This gel-like layer may also lead to a substantially prolonged dissolution of the miniature core after encapsulation due to the limited contact area with the solvent. The IM-HDT-WS took marginally longer to dissolve when compared to the Xioneer VXL130.

3.3.2 Over-moulding dissolution tests

The over-moulded lost-cores with the PMMA material as seen in Fig. 10a, b, were then subjected to the same dissolution conditions considered as in the cuboid samples. The time taken for the parts to achieve fully clean internal channels (without any remaining residual material) can be seen in Table 11.

Fig. 10
figure 10

Over-moulded Y-channels after dissolution of lost-cores, a Y0.8 and b Y1.6

Full size image
Table 11 Y-channel dissolution tests
Full size table

The over-moulding of the individual cores with the PMMA drastically reduced the effective contact area of the soluble core material with the solvent. This resulted in a decreased core dissolution rate when compared with the values obtained in the cuboid dissolution trials. During the dissolution of the cores from the over-moulded PMMA material, the AquaSys180 was the slowest in fully dissolving and also required the use of pressurised air and tools to mechanically aid in the removal of the residual gel-like material from the internal channels. This residual material which adhered to the internal walls of the channels may be attributed to the gel-like residual layer formation when the material is in contact with the solvent (tap water). The Xioneer VXL130 core was faster at dissolving from the over-moulded material. With regards to the IM-HDT-WS material, the dissolution of the material was faster than the AquaSys180 but was slightly slower when compared to the Xioneer (see Table 11).

4 Conclusion

The aim of this study was to investigate the feasibility of the manufacturing of components with meso–micro-scale internal features using the lost-core MIM process. This was accomplished through various tests conducted on the three investigated materials at the different stages of the process chain. The results from these trials/ tests revealed that:

  • The DMA tests revealed that the best candidate core material is the Xioneer VXL130, as it retained the highest value of the complex modulus across the 50 °C to the 110 °C range.

  • The DSC results revealed that the Tg value obtained for the IM-HDT-WS is 10.5% lower when compared to the manufacturer’s data sheet value. Further investigations are required to see whether this could be improved by optimising the post-curing process.

  • The DLP printed IM-HDT-WS cores had the best dimensional results with a difference of − 0.63% in the diameter and + 5% in the bifurcation angle of the Y1.6 channel and a difference of + 2.5% in the diameter and + 1.67% in the bifurcation angle of the Y0.8 channels.

  • The MIM over-moulding experiments showed that IM-HDT-WS also achieved the best dimensional results with a difference of 0% in the diameter and -3.17% in the bifurcation angle of the Y1.6 channel and a difference of + 4.88% in the diameter and + 11.48% in the bifurcation angle of the Y0.8 channels.

  • The dissolution tests showed that the Xioneer VXL 130 achieved a marginally faster dissolution rate to that of the IM-HDT-WS for both the Y1.6 and Y0.8 channels, respectively.

These initial results are encouraging since relatively good fidelity of the replicated core features can be achieved within the MIM process. However, further investigations are required to optimise the IM process, carry out a cost comparison of the proposed process chain with current manufacturing processes, and to study the feasibility of satisfactorily over-moulding 3DP cores with more complex internal geometries.