How do I choose the right material for my 3D printing project?

The short answer

Start from the failure you cannot accept

Materials are usually chosen by preference and then blamed for a failure. Inverting the order is faster: name the failure mode that would end the project, and the material follows from it.

• It must not deform in service - thermal, not strength, is the constraint. Anything above 80 °C leaves standard FDM plastics and points to PA12, polycarbonate, PEEK or printed metal.
• It must not shatter - toughness is the constraint, so PETG, ABS or TPU replaces PLA and standard resin.
• It must not creep or fatigue - the part is loaded continuously or cyclically, which points to SLS or MJF PA12 rather than any printed resin.
• It must not move out of tolerance - warping and shrinkage are the constraint, so material choice and build orientation matter as much as the process.
• It must not corrode or swell - chemical resistance decides it: PETG and nylon for mild chemistry, 316L stainless for aggressive service.

Writing that sentence down before the material is chosen removes most of the argument, because a part rarely fails for two reasons at once.

Step 1: service temperature

Ask for the steady temperature, the peak temperature and how long the peak lasts. Continuous exposure at 60 °C is a harder requirement than a two minute excursion at 90 °C, and a datasheet heat deflection number tells you neither on its own.

Read the bands as planning estimates: they mix heat deflection at 0.45 MPa with published maximum continuous use, and a loaded part fails at the bottom of each band rather than the top. The practical thresholds are that PLA and standard resin are indoor-only materials, PETG reaches about 75 °C, ABS and ASA reach about 100 °C, and PA12 from SLS or MJF sits around 170 °C. Above that, high-temperature resins and printed metal take over - 316L stainless is published for service above 300 °C. Also check the local heat: a housing next to a power supply sees a temperature the ambient specification never mentions.

Step 2: load and its direction

Then ask what the load is and which way it points. This is where process choice matters more than the polymer, because FDM parts are strongly anisotropic: they are built in layers, so they are weaker across the layer planes than along them. Published PLA data shows tensile strength around 35 to 40 MPa in the XY plane falling to about 31 MPa in Z, and the penalty grows with layer height and cooling.

Two design responses follow. Orient the part so that the main load runs along the layers rather than pulling them apart, and prefer SLS or MJF where the load comes from several directions at once, because powder bed parts are isotropic and have no support material to remove. Printed resin is isotropic too but brittle and poor in fatigue, so it suits static fit components rather than anything cycled. If the part is safety critical or carries multi-directional structural load, the honest answer is often printed metal or conventional machining rather than a different polymer.

Step 3: detail, finish and tolerance

Once temperature and load are satisfied, finish usually decides the process. Tolerance, surface roughness and minimum feature size travel together, and they are a property of the process more than the material.

FDM holds about plus or minus 0.5 mm with visible layer lines, SLA holds plus or minus 0.15 mm with a smooth surface, SLS and MJF sit around plus or minus 0.3 mm with a uniform matte finish, and SLM metal reaches plus or minus 0.1 mm before machining. If the drawing calls for tighter than plus or minus 0.1 mm, or for a surface finish below about Ra 1.6, the right move is to print for shape and finish the critical features by CNC rather than to keep hunting for a printer that will not get there.

Step 4: quantity and cost

Cost is last because it rarely changes which material is correct - it changes which process to use to make it. FDM is cheapest for one-offs and directional loads, SLA sits one tier up, SLS and MJF are priced for functional quantities, and metal printing is in a different category again.

The unit cost is driven by machine time, material, support or powder handling, and post-processing labour, not by the spool price. Five minutes of support removal at a 15 USD per hour shop rate adds about 1.25 USD to every part, which for a 1000 piece run is more than the material difference between two polymers. That is also why printing stops making sense at some volume: once the geometry is stable and the quantity is in the thousands, tooling for injection moulding or die casting usually beats any printing route on unit cost, even though it has a higher entry cost.

The four questions as a checklist

In practice the sequence closes fast. Temperature removes the unsuitable polymers, load direction decides between a directional process and an isotropic one, detail decides whether the process can hold the drawing, and quantity decides between printing and tooling. When the answers contradict each other - a tight tolerance with a high temperature, for instance - the resolution is usually a hybrid route: print or cast the form, then machine the critical features.

Material selection at SOMI Custom Parts

We run FDM, SLA, SLS, MJF, SLM and injection moulding, so the recommendation is not tied to one process. Send the model with the service temperature, the load case and the annual quantity and we will come back with the material, the process and the price, including a note when a moulded part is cheaper over your volume: see FDM 3D printing, SLS 3D printing or SLM metal 3D printing.