How the University of the Basque Country 3D prints personalized pills

• 1
  Key takeaways
• 2
  Why 3D prints personalized pills need a controlled filament first
• 3
  From powder blend to printed tablet: the four physical steps
• 4
  The engineering limits behind 3D prints personalized pills
• 5
  What printing a tablet shares with machining a medical part
• 6
  When printed tablets fit, and when they do not

A printed tablet is only as uniform as the filament that feeds it. The Basque Country work starts with starch-based material that is compounded with the active ingredient, then extruded into a filament of consistent diameter. If that diameter drifts, the deposited volume drifts with it, and the dose written into the file no longer matches the dose in the tablet.

This is the same logic we use on metal parts: control the stock, control the result. In CNC machining a bar of 6061-T6 has a known hardness and a known size, so a program cut for it repeats. In pharmaceutical printing the filament is the stock. Target diameter is usually held within roughly ±0.05 mm, and any segment outside that window is cut out before the spool reaches the printer.

Starch earns its place because it is cheap, it is already used as a pharmaceutical excipient, and it can be blended with a range of polymers to tune how the tablet behaves in water. It is also thermoplastic enough to be extruded at temperatures that many small-molecule drugs tolerate. Those three properties together are what make the process practical rather than theoretical.

The consequence is a narrow material window. A drug that degrades above the extrusion temperature cannot be printed this way, and a drug that needs an amorphous dispersion may need a different carrier altogether. The printer does not fix a bad formulation.

Step one is blending. The active pharmaceutical ingredient is mixed with starch and any plasticizer or release modifier until the powder is uniform. Poor blending shows up later as tablets that vary in potency from one to the next, and no amount of printer tuning recovers from it.

Step two is hot-melt extrusion into filament. The blend is heated and pushed through a die, then cooled and spooled. This is where the drug is dispersed in the polymer matrix, and where the thermal history of the batch is set. Cooling rate matters: fast cooling tends to lock in a more amorphous structure, slow cooling can let the drug recrystallize.

Step three is printing, usually fused deposition modeling. The nozzle lays down the filament layer by layer, and the infill percentage becomes a direct control on drug load per tablet. A 60 percent infill tablet carries less drug than a 90 percent infill tablet of the same outer dimensions, even though both look identical from the outside.

Step four is inspection and, if needed, coating. Weight, dimensions and dissolution profile are checked, and a film coat can be added to mask taste or delay release. Each step is individually simple. The difficulty lives in keeping them consistent, because a 3 percent drift in filament diameter or a 5 °C shift in nozzle temperature changes the dose.

Resolution is the first limit. A typical pharmaceutical FDM nozzle runs 0.2 mm to 0.4 mm. That sets the smallest feature you can print and the surface roughness of the tablet. It is coarse compared with the ±0.005 mm we hold on machined metal, but for a tablet the relevant tolerance is weight and dissolution, not surface finish.

Throughput is the second. Printing one tablet takes minutes, not milliseconds. That is acceptable when a hospital pharmacy prepares a small run for named patients, and unacceptable when a line needs to output millions of units per shift. The technology is a compounding tool, not a replacement for a rotary press.

Validation is the third and the hardest. In regulated markets every printed batch needs a documented link between the digital file, the filament lot and the finished dose. That means lot-level traceability of the API, records of extrusion parameters, and a release test per batch. Software control is convenient, but it also means the file itself becomes a controlled document.

Finally there is the release mechanism. Starch-based matrices typically swell and erode in the gut, so the drug comes out over a period set by the matrix rather than by a coating. Changing wall thickness or infill changes that period. This is useful for tailoring a profile, and it is also why the same file printed on a different machine may not release at the same rate.

A printed tablet and a machined implant solve different problems, but the quality logic is the same. Both start from a material lot with a certificate, both depend on a machine that holds its parameters, and both end with inspection that proves the part matches the drawing. Skip any of the three and the result is a part that passes once and fails the next run.

One difference is where the tolerance goes. On a machined component we hold ±0.005 mm on a bore because a mating part has to fit. On a tablet the critical dimension is mass, because mass carries the dose. Dimensional accuracy still matters, but only as a route to weight control.

For medical device work we hold ISO 13485:2016 and keep inspection records available on request. The same discipline applies to any drug-delivery hardware around a printed tablet: the printer frame, the nozzle holder, the spool hub. These are machined parts, and they need the same traceability as the tablet itself.

• 1
  Material lot controlDocumented properties for every incoming batch.
• 2
  Parameter lockNozzle temperature, feed rate and layer height recorded per run.
• 3
  InspectionWeight and dimensions checked, not assumed.