Curtin Researchers Trace 3D-Printed Ghost Guns to Filament Source

Researchers at Curtin University in Perth matched 26 of 28 printed PLA samples to their specific filament product, opening a possible forensic path to trace 3D-printed ghost guns. The study appeared on October 22 in the journal Forensic Science International.

The team, led by Dr. Georgina Sauzier of Curtin's School of Molecular and Life Sciences, worked with forensic experts from ChemCentre. Their method links a printed object back to its unprinted filament, down to the brand and type.

Why does this matter for law enforcement?

Conventional firearm investigations rely on serial numbers, registration records and licensed distributor networks. A 3D-printed plastic firearm has none of these. PLA parts are also compostable, which makes disposal trivial for the manufacturer.

The Curtin method could let police associate a seized ghost gun with filament seized elsewhere. That association would map a distribution network for printed firearms rather than a single weapon.

The problem is growing. Freely available design tools, including AI-generated 3D-printed gun models, have lowered the barrier to production worldwide.

How does the technique work?

3D printing alters the physical properties of the filament, and that alteration is measurable. The researchers used differential scanning calorimetry, or DSC, a thermoanalytical technique that sweeps the sample through a range of temperatures and records heat flow.

The instrument measures the power required to hold or change the sample's temperature, producing a graph of its thermal properties. That graph can be cross-referenced against the unprinted filament to establish an association.

PLA works because it is semi-crystalline. Its crystalline structures break down and rearrange at clear temperature transition points, which act as identifiers. The method does not yet work for other common print materials.

What are the limits?

PETG and ABS failed to produce identifiable signatures. Sauzier attributes the failure to their amorphous molecular structure, which yields very similar thermal results across products with no clear distinguishing markers.

The team also tested ATR-FTIR, attenuated total reflectance-Fourier transform infrared spectroscopy, which uses infrared light to identify chemical bonds and material composition. That method revealed differences in additives for ABS and PETG, but still fell short of linking a printed part to its filament product.

The research also carries methodological constraints:

• All samples were printed on a single machine.
• Test tiles were small and cut with scissors.
• Samples sat for a year in monitored office conditions before testing.

Cooling rates vary with print geometry and speed, and different printers may change the physical properties further. Each variable adds a link to the chain of uncertainty.

What comes next?

The researchers describe the work as early-stage and call for more rigorous testing. A 93 percent match rate on PLA is the starting point, not the finish.

If the technique matures to reliably identify a broad range of materials, it could trace a printed firearm to its printer and filament. That in turn could generate purchase records: where the filament was bought and who bought it. For now, the method covers one polymer family under controlled conditions, and validation across machines, materials and storage environments remains the decisive test.