How SLA rapid tooling is influencing medical device development

. Medical device product development includes the validation of the end-product before mass production can commence. This validation should be completed with a device that is manufactured from the final material and with the final manufacturing process. For plastic medical devices, Rapid Tooling can be used to manufacture parts for clinical validations at a cost-effective price. Further tooling cost reductions can be achieved if Stereolithography (SLA) is utilized to create injection moulding tools. This article determines the efficacy of SLA printed rapid tools to injection mould small plastic medical devices.


Introduction
Local and international regulatory bodies require a clinical evaluation to be carried out during the product development stage of any medical device [1]. The extent of this evaluation is dependent on the classification of the device which is largely dependent on the level of risk to the patient, user or public health. SAHPRA, the South African Health Products Regulatory Authority, classify medical devices from low-risk, non-invasive devices (class A), to high-risk implantable devices administering medicine (class D) [2]. The manufacturing and distribution of class A devices do not require licencing or approval from SAHPRA, whereas medium to high risk (class C) and class D devices has international pre-market approval as a pre-requisite.
SAHPRA approves a pre-market approval from the regulatory authority of Australia, Brazil, Canada, the European Union, Japan, USA, and the World Health Organisation. The U.S. Food and Drug administration (FDA) stipulates that a 510(k) Premarket Notification is based on substantial evidence showing conformance to a legally marketed device [3]. This necessitates the testing and validation of the product in the final material and process.
The common industry medical device product development process is depicted in Error! Reference source not found.a. For this process, if a medical device is to be clinically tested and validated, a production tool should be manufactured. This production tool entails a high initial capital cost which, when coupled with moulding challenges, can lead to excessive expenditure before a final part is manufactured.
Rapid Tooling (RT) is the manufacturing of injection mould tools with Additive manufacturing (AM) methods [4]. If RT is used as suggested in Error! Reference source not found.b, the moulding challenges can be resolved before high capital costs are incurred and cost-effective devices can be manufactured to be used during clinical testing and validation. Furthermore, RT allows for changes to be made to the mould and the manufacturing to take place in a shorter timespan compared to conventional subtractive manufacturing. Various metal and resin powder AM technologies, collectively known as Powder Bed Fusion (PBF), have been used to produce rapid tools. Direct Metal Laser Sintering (DMLS), Selective Laser Melting (SLM) and Selective Laser Sintering (SLS) are common PBF methods [5]. Rapid tool inserts manufactured from Maraging steel has shown to be a good candidate for injection moulding mass production runs, however, manufacturing costs are high when compared to subtractive manufacturing processes [6]. SLS inserts manufactured from polymer-metal composites such as Alumide ® have shown to be a viable alternative to DMLS maraging steel inserts where injected polyamide temperatures do not exceed 230˚C [7][8][9].
Another AM technology used for RT is Material Jetting (MJ) which has been shown to produce injected polyamide parts with high dimensional accuracy [10][11][12]. The disadvantages of MJ tools are the high cost associated with materials and machine maintenance as the different nozzles can get clogged during use [13].
Rapid tools have also been manufactured with Stereolithography (SLA) which showed the successful moulding of plastic parts albeit in low quantities [14][15][16]. SLA rapid tools are becoming increasingly cost-effective as desktop 3D printers are becoming readily available [17]. Moreover, SLA rapid tools exhibit a smooth surface finish when compared to SLS and DMLS mould inserts without surface post-processing [7,17].
The plastic injection moulding of medical devices with SLA rapid tools is, however, not documented in literature. With clinical testing and validation only requiring a low number of parts, SLA IM tools could be deemed the most cost-effective RT method to produce plastic medical devices. No cooling channels, conventional or conformal, were added to the mould inserts to ensure that the mould inserts were not weakened at any point.

Manufacturing of tooling
The printer used to produce the mould inserts was a Formlabs Form 3B. This printer can produce parts from various photopolymer resins with a high surface finish quality.
The mould inserts were grown using Rigid 10K resin (FLRG1001) from Formlabs. The mould inserts were placed directly on the build platform to remove the need for supports and to ensure that the bottom surface was flat.
The grown mould inserts had a noticeable concave shape in the centre of the top surface (see Fig. 2). This concave shape would result in excessive flashing in the centre of the part. The risk of excessive flashing was mitigated by manufacturing the part 500µm higher, with a straight extrusion added, and to machine the excess material off to create a flat surface.

Post-processing of mould inserts
After printing, the mould inserts were washed in a Form Wash for 20 minutes and left to air dry. After drying, the inserts were cured in a Form Cure device as per Formlabs UV curing specifications (70˚C for 60 minutes) [18].
The parts were machined in a high speed 3-axis milling machine. An indexable 50 mm diameter face mill with all-purpose inserts was used to remove 250µm material with two cuts. The machine feed rate and cutter speed were set to the manufacturer's specified limits. The resulting insert had a flat top surface which showed excessive damage to the cavity features (see Error! Reference source not found.).
Notice no gap between mould inserts on front face while gap appears at the back of the gate closer to the centre of the mould Following these results, it was decided to use an indexable 32mm chase mill with aluminium inserts to remove the excess material (see Fig. 4). This decision was based on the means with which the cutter removes material and chips. The aluminium insert has a high rake angle and sharp cutting edge to remove glutinous material while also affecting a low cutting load on the mould inserts. The results following cutting with the Aluminium inserts showed less damage to the cavity features compared to the all-purpose inserts but still significant damage when compared to metal milled mould inserts.
Lastly, a 16mm carbide end mill was used to further decrease the cutting load on the mould inserts. This process resulted in the lowest damage to the mould inserts (See Fig. 5).

Fig. 5. Location of edge breaks after milling insert with end mill
Mould inserts in their uncured or green form were also machined to evaluate the effect of curing on part machinability. The results showed that uncured mould inserts produced the same level of damage to the cavity edge (see Fig. 6 and Fig. 3). The green parts had the added risk of damage during clamping in the mill. The effect of machining on the insert face did not provide the desired outcome i.e., a split line similar to a metal insert. Ultimately it was decided to not perform any post-processing on the insert and rather grow the moving side insert 25µm higher and allow the clamping pressure of the injection moulding machine to deform the concave insert face into a flat surface.

Material used for injection moulding
Three thermoplastic materials were selected to be used to manufacture the device. Polypropylene (PP), Polyoxymethylene or Acetal (POM) and Polybutylene Terephthalate (PBT). The PP and POM materials was mixed with a radiopaque additive to evaluate the final strength of the part.
The additive was mixed with the thermoplastic pellets in powder form, but the first off tool samples showed inconsistency in thermoplastic-additive binding. Following these results, a premixed masterbatch in pellet form of the additive and both PP and POM as base material was used.

Experimental results
The moulding of the parts was carried out with an Arburg Allrounder 470C (1500kN) injection moulding machine. The machine settings such as injection pressure, injection speed, shot size etc., were determined by moulding off-tool samples until a machine steady-state and good part repeatability was achieved. The heating zone temperatures were ascertained from the material specification sheet supplied by the manufacturer.
The thermoplastics and masterbatch were mixed and dried prior to moulding according to manufacturer specifications. The thermoplastics were not handled on the same day to decrease the risk of material cross-contamination.

Polypropylene and Polyoxymethylene parts
The PP parts were moulded with a nozzle temperature of 200˚C, an injection pressure of 80 bar and an injection speed of 166.7mm/s. A total of 125 parts were moulded until mould failure.
The POM parts were moulded with a nozzle temperature of 210˚C, an injection pressure of 140 bar and an injection speed of 85mm/s. A total of 13 parts were moulded until mould failure.
During the moulding of both the PP and POM parts, the surface of the SLA mould inserts and especially the part cavities was cooled by blowing compressed air across it. A medical grade mould release was subsequently applied to the part cavities to decrease the risk of part adhesion to the mould inserts. This process was repeated after the moulding of ten parts.

Polybutylene Terephthalate parts
The PBT parts were moulded with a nozzle temperature of 265˚C, an injection pressure of 115 bar and an injection speed of 100mm/s. A total of 134 parts were moulded until mould degradation to a point where part flashing was exceeding acceptable limits. The PBT mould inserts did not show any signs of cracks or failure as with the PP and POM mould inserts.
The resulting PBT parts, moulded from inserts with no post-processing, displayed a split line similar to that of a part moulded with a metal insert (See Fig. 7). When comparing the original part dimensions from the Computer Aided Design (CAD) with part number 100, the height of the part measured 14µm undersize. The moulding of the PBT parts did not include the cooling process followed during the moulding of the PP and POM parts. This cooling process was neglected as the PBT material would solidify and create a nozzle plug after moulding two parts. This necessitated the mould inserts to be exposed to high moulding temperatures with no cooling applied between shots.

Maraging steel mould inserts
The maraging steel inserts are to be grown by Metal Heart with an SLM Solutions 280 HL system. At the time of publication, the inserts have not been grown but a quote has been supplied by Metal Heart to estimate the costs involved. Table 1 displays the cost of printing the moulds as well as the cost of the post-processing. The cost of pins, pin-retainer and support plates and injection moulding machine hours are neglected. The printing cost shown is for a complete mould insert set consisting of a fixed and moving side.

Conclusion
This study determined the efficacy of SLA manufactured rapid tooling inserts for the manufacturing of a small plastic medical device. The process of manufacturing the inserts as well as the post-processing processes implemented to increase part quality were discussed.
The small plastic medical device selected for this article necessitated a rapid tool since the part dimensions was not manufacturable with Computer Numerical Control (CNC) milling techniques. A direct comparison with subtractive manufactured mould inserts could therefore not be made as an Electrical Discharge Machining (EDM) manufacturing method would have to be employed. However, a mass-production injection moulding machine was utilized to expose the SLA mould inserts to as close to an industry standard manufacturing environment as possible. The injection moulding process followed throughout the IM trials can thus be applied to any SAHPRA licenced or ISO 13485 accredited medical device manufacturer.
The SLA RT inserts were found to be a cost-effective alternative to maraging steel RT inserts for the manufacturing of a small batch for clinical evaluation. Moreover, the SLA inserts enabled the cost-effective testing of the part-and tool design and the effect of the radiopaque additive on the moulding process. This testing led to the mitigation of three manufacturing risks and produced parts to a quality suitable for clinical evaluation.