Fused Filament Fabrication, Stereolithography and Silicone Direct Ink Writing: A Multi-Property Benchmark on a Common Specimen Set

Published comparisons of polymer additive manufacturing (AM) processes are usually assembled from different laboratories, geometries and test protocols, rarely include elastomers, and seldom report durability, biological response and environmental burden next to geometry and strength. Engineers therefore lack a consistent basis for choosing a process–material system for parts that combine rigid and compliant functions, such as soft-robotic grippers, housings with integrated seals, wearable devices and laboratory hardware. We benchmarked five systems—material extrusion (MEX) by fused filament fabrication (FFF) of PLA and ABS, vat photopolymerisation (VPP) by stereolithography (SLA) of Clear V4 and Tough 2000 resins, and extrusion-based direct ink writing (DIW) of a platinum-cure silicone (Ecoflex 00-30)—on 228 specimens made and tested with one protocol in one laboratory. Pairing two materials within each rigid process separated process-controlled from material-controlled properties: surface finish, dimensional accuracy, porosity, electricity, cost and cytotoxicity were governed by the process, whereas strength, ductility, fatigue ratio and ageing resistance were governed by the material. SLA gave the smoothest surfaces (Ra 1.2–1.6 μm, against 8.4–11.1 μm for FFF and 3.8–5.1 μm for DIW) and mean deviations of at most 0.052 mm on 20 mm cubes. Normalised by strength, the rigid materials had similar fatigue behaviour (106-cycle fatigue ratios 0.34–0.42). Exposure to 500 h of UV reduced the strength of every material (82.7–91.0% retention). Both SLA resins fell below the 70% cell-viability threshold of ISO 10993-5, and SLA bars carried the largest carbon footprint (141–170 g CO2-eq per bar, against 54–94 g), an order that held for grid carbon intensities up to 800 g CO2-eq kWh−1. In an exploratory comparison of rigid–soft hybrid routes, integrating DIW into the FFF build shortened cycle time by 25–32%, and none of its six specimens failed adhesively. A TOPSIS analysis of 12 commensurable criteria, with criterion and weight sensitivity, showed that rankings across material classes depend on which criteria are included, so selection should apply class constraints first and rank within the admissible class second.