3D printing stainless steel with metal filament: how viable is it
Bound metal deposition makes metal parts with a filament printer: a polymer loaded with metal powder is printed, the polymer is removed, and what remains is sintered. The equipment costs a fraction of a powder-bed fusion system. This thesis implemented the full chain for 316L stainless steel and assessed how far it goes. The main result is that the porosity of the final part is set during printing, and sintering does not correct it.
Context
In Colombia, a metal spare part with complex geometry is usually imported, with weeks of waiting. Fusion-based metal additive manufacturing solves the technical problem but demands an investment and infrastructure that few companies in the country can take on. The thesis included a survey of companies in the sector and took the impeller of a centrifugal pump as its case study.
The process has three stages. The printed part, or "green" part, contains metal and binder. Debinding removes the polymer. In sintering, at a temperature close to melting, the metal particles bond and the part shrinks.
Methodology
Printing. A filament printer was set up to work with 316L filament, and cylindrical specimens and the impeller were made.
Specimen model and deposition paths
Printing parameters were studied with a six-factor Taguchi design: extrusion speed, temperature, layer height, line width, extrusion factor and infill-perimeter overlap. Surface quality was quantified with a vision algorithm that detects the contour and the defects in scanned images of each specimen.
Image processing to quantify surface defects
Main effects of the six factors on the surface defect index
Sintering. Two routes were compared: a conventional resistive furnace and induction heating. In both, the part is buried in an alumina and graphite ballast inside a crucible, covered with activated carbon.
Alumina crucible setup for induction sintering
Specimens after induction sintering at different target temperatures
Characterization. Metallography, porosity measured by image analysis, hardness and chemical composition.
Results
Porosity is born in printing. Voids between beads that are not close enough to each other are preserved through the whole chain. Sintering consolidates the structure it receives and does not close those defects. Taguchi optimization improved the surface of the green parts without improving internal density to the same degree. Good surface quality does not guarantee integrity inside.
Induction densifies the surface, not the core. With induction, heating concentrates at the periphery of the part.
Metallographs of the perimeter (a) and the core (b) of a specimen induction-sintered at 1250 °C
Average porosity at perimeter and core against process temperature
As temperature rose from 1100 to 1250 °C, perimeter porosity fell from 8.4% to 0.8%, while core porosity stayed between 18% and 25%. The result is a part with a dense shell and a porous interior, with a matching hardness gradient. The conventional furnace gave more uniform consolidation.
The chemistry changes. Composition analysis showed an increase in carbon and a reduction of chromium in solution, with carbide precipitation at grain boundaries. The cause is carbon transfer from the support medium during the hold at high temperature. In a stainless steel this compromises corrosion resistance.
What is missing
The initial hypothesis was to reach relative densities above 95%, and it was not met through the volume of the parts. Mechanical characterization relied on hardness, so performance under load is inferred from it and from porosity. A toolpath design aimed at internal density and a support medium that does not supply carbon remain to be explored.
The thesis's conclusion is bounded: the technology is viable for prototypes, auxiliary parts and components under moderate loads, provided the design accounts for porosity. It does not replace fusion processes in demanding structural applications.
How it fits in Robiolab
This is not a robotics project. It belongs to the digital-fabrication line the group has worked on since its beginnings, with the stereolithography printer and the electrospinning printer. Its relevance to the lab is practical: the ability to make small, complex metal parts locally, such as gears, couplings or mechanism structures, is a recurring limitation when building prototypes of robots and wearable devices. The dense-shell, porous-core structure that induction produces is a defect for a spare part, but it is also a property gradient that could be designed on purpose.