How to Choose Polymer Materials for 3D Printing

• 1
  Key takeaways
• 2
  Define the service conditions before you look at any polymer materials for 3D printing
• 3
  Match the process to the geometry, then the polymer
• 4
  Read the data sheet the way an engineer reads it
• 5
  Set the print parameters that the material actually needs
• 6
  Know when 3D printing is the wrong answer
• 7
  Step by step: how to choose polymer materials for 3D printing
• 8
  Polymer families for 3D printing compared

Most bad material choices are made in the first ten minutes, before anyone writes down what the part has to survive. Start with four numbers: maximum continuous service temperature, peak load and its direction, chemical exposure, and required dimensional tolerance. Write them down. Then compare candidates against those numbers, not against marketing data.

Service temperature is the fastest filter. PLA softens near 60 °C, PETG holds to roughly 70 °C, ABS and ASA to about 90 °C, and PC or PA blends push past 110 °C. PEEK and PEI are used above 150 °C. If your part sits near a motor, a lamp housing or an engine bay, PLA is out before you open a catalog.

Load direction matters as much as load size. An FDM part is a stack of welded beads, so the bond between layers is the weak plane. A bracket printed flat and loaded in its own plane can carry two to three times the load of the same bracket printed upright. Decide the build orientation on paper while you are still choosing resin or filament.

Chemical exposure is often forgotten. ABS dislikes acetone and many solvents, PP resists most of them, and PA absorbs moisture and swells. If the part will sit in cutting fluid, IPA or hydraulic oil, check compatibility before you commit to a print run.

• 1
  Write the four numbers firstTemperature, load, chemistry, tolerance. Everything else is secondary.
• 2
  Reserve 20 °C of headroomA material rated at 90 °C should not run continuously at 88 °C.
• 3
  Note the load planeMark it on the drawing so the print orientation follows it.

Process and material are a package. FDM gives you the widest material range and the lowest cost per part, but it leaves visible layer lines and a rough as-printed surface around Ra 8–16 μm. SLA and DLP give smooth surfaces and fine features down to 0.1 mm, but the resin palette is narrower and parts are brittle until post-cured.

SLS and MJF use powder and a laser or fusing agent. Parts come out nearly isotropic, need no support structures, and handle internal channels and nested geometry that FDM cannot. The trade-off is surface finish, which stays grainy and usually needs bead blasting.

Material extrusion and vat polymerization both shrink. Expect 0.3–0.8 percent linear shrink on FDM in ABS and up to 2 percent on some resins. SLS nylon shrinks more but predictably, and the machine compensates. If your tolerance is tighter than ±0.1 mm over 100 mm, plan for a finishing cut or move the part to CNC.

A useful rule: pick the process that can hold your tightest feature, then pick the polymer that survives the environment. Doing it the other way around usually ends with a material you cannot print at the resolution you need.

• 1
  FDMWidest material choice, visible layers, anisotropy to manage.
• 2
  SLA / DLPBest detail and surface, limited toughness, post-cure required.
• 3
  SLS / MJFNear-isotropic, support-free, grainy surface.

Printed test bars and molded test bars are not the same object. Most published tensile numbers for filaments come from injection-molded specimens, which have no layer boundaries and no voids. A printed part in the same polymer typically reaches 50 to 80 percent of that value, and less across the layers.

Look at four lines on the sheet: tensile strength, tensile modulus, heat deflection temperature at 0.45 MPa, and elongation at break. Elongation tells you whether the material will crack or yield. Under 5 percent, expect brittle failure with little warning. Above 50 percent, the part will bend and stay bent.

Glass transition temperature is not a service limit. A polymer can be well below its Tg and still creep under a constant load. For anything that stays bolted down and loaded for months, creep data matters more than a single tensile number.

Moisture is the quiet killer. PA and PC absorb water from the air and must be dried before printing, typically 4 to 6 hours at 70–80 °C for PA and 3 to 4 hours at 80 °C for PC. Print wet filament and you get bubbles, poor layer bonding and a part that fails at half the expected load.

• 1
  Discount molded dataAssume 50–80 percent of it in a printed part.
• 2
  Elongation predicts failure modeBrittle or ductile, before you print anything.
• 3
  Dry hygroscopic polymersPA and PC need a drying cycle, not a guess.

Once the polymer is chosen, the parameters are mostly determined. Nozzle temperature, bed temperature, chamber temperature, layer height and wall count each have a working band, and stepping outside it changes the part more than switching brands does.

Layer height controls the trade between speed and strength. A 0.2 mm layer is a good default for functional parts. Going to 0.1 mm improves surface and Z-strength slightly but roughly doubles the print time. Going to 0.3 mm saves time and costs bond strength.

Wall count matters more than infill for stiffness. Four to six perimeters at 0.4 mm nozzle diameter carry most of the load in a bracket. Infill above 40 percent adds weight and time without much gain, because the shell is already taking the bending stress.

For high-temperature polymers, the chamber is the parameter people skip. Printing PEEK or PEI without a heated chamber above 150 °C produces warping and delamination no matter how good the filament is. If you cannot control the chamber, choose a different polymer.

• 1
  0.2 mm layerDefault for functional parts; go finer only for cosmetics.
• 2
  4–6 perimetersBeats high infill for stiffness in bending.
• 3
  Chamber temperatureNon-negotiable for PEEK and PEI.

3D printing wins on low volume, complex geometry and fast iteration. It loses on tight tolerance, high load and long-term dimensional stability. A printed bracket that carries 200 N in a lab may creep to failure after six months of constant load at 60 °C.

If the part needs ±0.05 mm or better, or a surface below Ra 1.6 μm, printing will not get there without secondary machining. At that point the honest comparison is printed-plus-machined against machined from stock. For small runs, machining from POM, PA or PEEK bar stock is usually faster and more predictable.

The same logic applies to press fits, threads and bearing bores. Printed threads strip easily. A printed bore for a bearing is rarely round enough. Design the print with 0.5 mm of stock on those features, then ream or bore them on a CNC after printing.

The practical split: use printing for housings, covers, ducts, jigs and form-and-fit prototypes. Use CNC for anything with a tolerance callout, a thread, a bearing seat or a sustained load. Many projects need both, and running them in parallel saves a week.

At GreatLight we run 127 CNC machines across three plants in Dongguan and Singapore, so a printed prototype can move straight into a machined or cast production part without changing suppliers. The same DFM review covers both routes.

• 1
  Print for form and fitHousings, covers, ducts, jigs, early prototypes.
• 2
  Machine for tolerance and loadThreads, bearing bores, press fits, sustained stress.
• 3
  Hybrid is normalPrint first, then machine the critical features.