How to choose a material for a 3D printed part that must withstand heat and load

Choose the material by the highest temperature including peaks, the type and duration of the load, and the environment. The manufacturer's heat resistance figure (PETG 68 °C, ASA 93 °C, PC Blend 113 °C) is the heat deflection temperature at low load. It compares materials and is not a safe continuous temperature under load, because plastics slowly creep.

This is how we propose materials for custom 3D printing. This text is a selection method for heat and load. Materials for end-use parts is an overview by type of load, and the basic comparison describes the individual filaments. The numbers below are figures from Prusa Polymers for its branded Prusament filaments, and other brands may differ.

What the heat resistance figure for PETG, ASA and PLA actually means

Catalogues mix up three different things:

• Glass transition temperature (Tg). For amorphous plastics such as PLA and PETG it is the change from a hard to a soft state, not melting (Polymer Science Learning Center). An independent study found 69 °C for PLA and 73.5 °C for PETG (Martins 2024).
• HDT (heat deflection temperature, ISO 75). A bar under a set load is heated at 120 °C per hour, and HDT is the temperature at which it deflects by 0.20 %. A manufacturer of testing machines calls it a comparison value, not the maximum service temperature of a finished product (ZwickRoell).
• Vicat (ISO 306). The temperature at which a needle penetrates the sample to a depth of 1 mm (Formlabs).

For these materials Prusa does not use the abbreviation HDT on its website, but its figures of 55, 68, 93 and 113 °C match HDT at 0.45 MPa in the technical data sheets. That match is our interpretation. For PA11 CF it is 192 °C, and the website states up to 190 °C depending on load. For PLA and PETG, HDT sits below the glass transition temperature (55 against 69 °C, 68 against 73.5 °C), so the limit for a part is lower than the temperature at which the plastic softens.

Why 68 °C does not mean safe under load

Prusa tests HDT at two loads, 0.45 and 1.80 MPa. For PLA and PETG the values are the same, and for the others they fall at the higher load:

Sources: technical data sheets (PLA, PETG, ASA, PC Blend, PA11 CF), specimens printed on Prusa printers at 100 % infill. We calculated the differences ourselves.

At the higher load PC Blend is only 7 °C ahead of ASA. Unfilled nylon fares even worse: for the unfilled PA11 pellets Rilsan FMNO (not a filament, just an illustration) HDT is 129 °C at 0.45 MPa and 42 °C at 1.8 MPa (Arkema). “Nylon withstands 190 °C” therefore holds only for carbon-filled types, and we always agree the exact type of nylon in advance.

ASTM D648, the American counterpart of ISO 75, states explicitly that these data are not intended for design or for estimating endurance at elevated temperature (ASTM).

Creep and fatigue: a part gives way below the limit

Plastic under constant load creeps slowly, even when the temperature stays below the limit. In a peer-reviewed study (Ultimaker materials, temperatures of 25, 40 and 60 °C, stress of 10 and 20 MPa, three hours) load had a stronger effect than temperature and PLA fared worst. ABS broke in about three minutes at 60 °C and 20 MPa, while polycarbonate survived every combination (Dogan 2022). The specimens were machined, the test was short and it did not include PETG or ASA, so we take it as a direction, not a table of limits.

Repeated loading is stricter. After a million cycles the fatigue limit was 25 % of static strength for ASA, 17 % for nylon 12 and only 7 % for polycarbonate (an industrial Stratasys machine, Martins 2026). The numbers belong to another machine and other materials, but they show that under cyclic load you cannot size a part by the catalogue tensile strength.

A five-step selection method

If failure of a part could endanger life, health or property, it is not an ordinary material choice but a matter for a special written agreement (terms and conditions). For other parts we proceed like this:

1. Highest temperature including peaks, and for how long. Add up the surroundings, self-heating (power supply, motor) and sun. A car cabin in direct sun without ventilation can exceed 70 °C (Grundstein 2010, US measurements), so PETG at 68 °C is not enough. Compare it with HDT at the matching load. We did not find a safety margin for HDT in a standard, so we apply our own and test a sample on an important part.
2. Type and duration of load. Impact, constant, repeated, compression, bending or shear, and the direction of the force relative to the layers. Under constant load a thicker cross-section and lower stress help.
3. Environment. UV, moisture, oils, solvents, cleaning agents, hot surfaces and friction that heats the part.
4. Material from the table below.
5. Design. Orientation, perimeters, inserts and wall thickness, more below.

Table: heat and load, material, reason and risk

The temperature bands are our grouping based on the data above, not guaranteed limits.

Layer direction, perimeters and infill belong to the material choice

An FDM part is weakest between the layers. Prusa gives an interlayer adhesion of 11 MPa for ASA and 18 MPa for PETG, while the strength of a specimen in the layer plane is 42 and 47 MPa. The data sheet does not describe the measuring method, so only as a rough guide: the bond between layers is about a quarter to a half of the in-plane strength. For tension we therefore orient the layers parallel to the force and lay pins flat (Hubs).

Strength is determined mainly by the number of perimeters, and infill helps mostly in compression (Prusa Knowledge Base). In a test with PETG at 20 % infill, the maximum tensile force rose from 559 N with one perimeter to 739, 913, 1071 and 1207 N with two to five (Menargues 2025). Infill helped too: with two perimeters the load capacity of PETG rose almost linearly as infill went from 20 to 80 %, by 813 N in total. We would rather anchor a screw in a threaded insert than directly in the plastic, and the print parameters are on the FDM printing page.

Annealing, moisture and chemistry: what changes heat resistance

Annealing in an oven looks like a cheap fix for PLA and PETG. In Prusa's test annealing worked best at 90 °C for PLA and 110 °C for PETG, yet on a real extruder cover it judged annealed PLA and PETG unusable for complex precise shapes, and ASA and ABS unsuitable for annealing (Prusa). In a study annealing raised the strength of PLA by 6 % and PETG by 8 %, but PETG changed length by about a tenth with unsuitable parameters, and the authors recommend a smaller layer height for strength instead (Stojković 2023). For heat we therefore choose a different material.

Nylon absorbs water and the type decides: in an independent, non-peer-reviewed test, carbon-filled PA6 kept 56 % of its strength and about a third of its stiffness after conditioning, while carbon-filled PA12 lost around 15 % of its strength (CNC Kitchen). These are different nylons from Prusa's PA11, so treat the number as a warning only. For oils and cleaning agents we confirm the choice with a test on a sample.

How we handle it on a job

In custom 3D printing we ask five things before proposing a material: the highest temperature including peaks, where the force comes from and whether it is constant, repeated or impact, the environment, the number of pieces and whether the part is load-bearing. From the answers we propose the material, orientation, perimeters and inserts. We print PC Blend and filled materials on request.

For an important part we recommend a test piece in real conditions before small-batch production starts or spare parts go into service. The printability check is free, we send a quote within 24 hours, and rates are in the price list. Tell us what the part has to handle.