Why 3D Printed Parts Fail Around Oil, Fuel and Solvents

The part sat in the workshop for a year without complaint. Then it spent an afternoon next to a parts washer, or you wiped it down with isopropyl before painting it, and a week later it is in two pieces. 3D printed parts fail around oil, fuel and solvents in four distinct ways, and the useful thing is that each one leaves different evidence. A part that went soft and a part that cracked clean were killed by different mechanisms, and swapping filament fixes only one of them.

This post contains affiliate links. If you purchase through these links, 3D Prototyping Hub may earn a small commission at no extra cost to you.

This post also contains Amazon affiliate links. As an Amazon Associate I earn from qualifying purchases.

Read the Failure Before You Change Material

Three variables decide all of this, and people usually record only the first: the chemical, the temperature, and whether the part is under stress. Temperature is a multiplier — a material that tolerates a fluid indefinitely at 20°C can swell in it at 60°C. Stress is a switch: the third failure above cannot happen at all in an unloaded part, and happens readily in a bolted one.

Stress Cracking Is the One That Catches People

Dissolution is obvious. You can watch it happen, and nobody misdiagnoses a part that went sticky.

Environmental stress cracking is the opposite, and it is responsible for most of the "but that chemical is supposed to be compatible" failures. It needs three things at once: a susceptible polymer, a fluid that does not dissolve it, and tensile stress. The fluid migrates into surface flaws exactly where the material is in tension and lets a crack open at a load the dry part carried without noticing.

The tension does not have to come from anything you did. A bolt torqued into a boss, a press-fit pin, a snap arm held closed, a part clamped flat against a surface it does not quite match — all of them qualify, and so does the frozen-in stress left over from printing, which is why a tall part can crack with no external load at all.

The combination to know on a desktop machine is isopropyl alcohol or a strong cleaner against PETG or polycarbonate. Both materials are listed as fine with alcohols, and both of them are, unstressed. Wipe down a bolted PETG enclosure with IPA and you may find it cracked at the screw bosses a fortnight later. Soap and water carries none of that risk.

What Each Material Actually Tolerates

Nylon is the hydrocarbon material. Petrol, diesel, engine oil, gearbox oil and grease do essentially nothing to it, which is why moulded nylon turns up throughout fuel and lubrication systems. The catch is the opposite of a chemical problem: nylon absorbs water out of ordinary room air within hours, and a wet spool prints a foamy, porous, weak wall — precisely what a chemical-contact part cannot have. Plain nylon filament needs a dryer beside the machine, not just dry storage, and PA6-CF trades some of that moisture sensitivity and warping for a hardened-nozzle requirement. Our nylon filament guide covers where the grades differ.

Polypropylene is the chemistry material. It is what laboratory bottles, chemical tanks and battery cases are made of, and against acids, alkalis, bleach and alcohols nothing else on a desktop comes close. It is also the hardest common filament to print: it shrinks, it warps, and it will not bond to a normal build surface — most people run PP filament onto packing tape of the same polymer, because like sticks to like.

Polycarbonate is the trap, for the second time. It is the strongest and most heat-tolerant material most people can print, and it is near the bottom of this table. Acetone attacks it, alkalis hydrolyse it, and it is one of the most stress-crack-prone engineering plastics in existence. Strength charts and chemical charts rank materials in almost opposite orders, and choosing PC for its numbers is how a part ends up cracking in a cleaning cupboard.

PETG is fine until something is holding it. Against water, dilute household chemistry, oils and greases it does well, and for an unstressed tray, funnel or splash guard generic PETG is a sensible, cheap answer — see the PETG buyer's guide for brand-level differences. Bolt it down and wipe it with solvent and you are in the previous section.

The Layer Seam Is a Path, Not a Wall

Even with the right polymer, an FDM wall is not a moulded wall. Every pair of adjacent beads meets along a continuous spiral seam that runs from the inside of the part to the outside, and a liquid that cannot attack the plastic will still travel along it.

That produces a failure that looks like the material's fault and is not. The fluid wicks into the wall, reaches the heat-set insert, the bonded joint, the electronics or the steel part on the other side, and the damage shows up a long way from where the liquid touched it. Chemical contact makes the case for dense walls much stronger than ordinary watertightness does, and the full version of that argument — perimeters, overlap, nozzle size, temperature — is in why printed parts leak. A 0.6mm hardened nozzle lays a fatter, hotter bead that welds to its neighbour more completely, and it is the hardened version because filled nylon widens a brass bore within a spool or two.

Fix It in This Order

1. Write down the three variables. The chemical, the peak temperature it reaches, and whether the part is in tension. Most compatibility charts are quoted at room temperature and no load, and both assumptions are usually false in a workshop.

2. Match the material to the chemistry, not to the strength chart. Hydrocarbons mean nylon. Acids, alkalis and aqueous cleaning mean polypropylene. Mild and unstressed means PETG. Nothing means PLA.

3. Design the seal instead of trusting the plastic. A joint that has to hold liquid should be an O-ring squeezed in a groove, with the plastic doing nothing but holding it. Elastomer seals survive chemistry and movement that destroy a bonded seam.

4. Treat fasteners and adhesives as part of the chemical system. Cyanoacrylate embrittles with age and disappears on contact with acetone — a joint can fail while the parts either side are untouched, which is covered in why glued prints come apart. Use a two-part epoxy where a joint meets solvent, and heat-set inserts rather than screws cutting their own threads — a screw biting into plastic builds a ring of tensile stress exactly where a solvent wants to start a crack, and stripped printed threads are the other half of that story.

5. Reduce the stress if you cannot change the material. Slotted holes rather than round ones, washers under bolt heads, a torque you can actually repeat, generous fillets at every internal corner, and annealing to relieve frozen-in print stress. None of this changes what the chemical does to the polymer; all of it removes the tension stress cracking requires.

6. Test it, because 72 hours answers the question. Print three identical coupons. Weigh one on a jeweller's scale, soak it in the actual fluid at the actual temperature for three days, and weigh it again: more than one or two percent of weight gain means the fluid is going into the material. Bend a second one against a former and soak it loaded — that is the stress-cracking test, and it fails parts that pass the unloaded one. Keep the third dry as a reference.

The Hardware These Materials Ask For

Nylon, PP and PC all want a hotend that holds temperature over 260°C under sustained flow, and warm, still air around the part — in a draughty room they shrink unevenly and split along a layer line partway up. The specification that decides it is a published chamber temperature, not the word "enclosed" printed on a box: a box with a lid and a box with a heater are different machines. QIDI's enclosed machines publish a chamber figure alongside the hotend rating.

Nylon also brings two consumables with it. A heated dryer box that feeds the printer directly, because a spool left out overnight is wet by morning; and a garolite build surface, which nylon grips when PEI will not hold it.

If PETG-grade duty is genuinely all you need, none of that applies — the relevant specifications are the hotend's maximum temperature and whether a 0.6mm nozzle is offered for your model, both of which ELEGOO's US store lists with the machines.

Where People Create the Problem Themselves

Most chemical damage in a home workshop is self-inflicted during cleaning and finishing.

• Do not degrease a stressed PETG or PC part with alcohol. Soap and water for anything bolted together. Save the 99% isopropyl for resin washing and for wiping a surface before bonding, where the part is loose and goes nowhere near a load.
• Keep solvent containers closed. Acetone vapour from an open washer or a rag bin hazes PLA and PETG on a nearby shelf and crazes anything stressed, without ever touching the liquid.
• Know what your support solvent attacks. D-limonene dissolves HIPS supports beautifully and also softens ABS slowly, so a long soak costs you surface detail on the model as well as the supports.
• Washed resin parts are already carrying solvent. Cured photopolymer absorbs isopropyl during washing and embrittles as it dries, which is why an under-washed or over-cured resin print snaps in the hand weeks later — the sequence is in resin safety and post-processing basics.

What Does Not Work

• More infill. The chemical meets the surface and travels along the seams. Solid infill adds mass, cost and internal stress — and internal stress makes stress cracking more likely, not less.
• A coat of paint or lacquer. Thin films are microporous and crack where the substrate moves. An epoxy lining brushed on thick is a real seepage fix in a tray or tank; a rattle-can clear coat is not a chemical barrier.
• Choosing the stronger material. PC beats PP comfortably on every mechanical number and loses to it badly in a cleaning cupboard.
• PTFE tape over the wrong polymer. Sealing a threaded joint does nothing about a wall that is swelling.

When to Stop Printing It

There is a line here worth naming. A part that holds fuel, runs hot in oil, sits in a pressurised circuit or has a safety consequence if it splits is not a desktop FDM part, however good the filament choice is — the seam is a real property of the process and no setting removes it. The sensible routes are moulded or machined stock polymer (PP, acetal, PTFE, UHMW), or SLS nylon, which has no spiral seam and near-isotropic strength. High-temperature printing services covers the bureau side of the same material question.

That is the crossover to get a quote on rather than another spool. The providers in our directory quote machining, moulding and industrial printing side by side, and the background on deciding between them is in our guide to functional and end-use printed parts.

Print the prototype in nylon, confirm the geometry, soak a coupon for three days, and let the failed coupon rather than the failed assembly tell you whether the material was right.

Hero photograph by Snapmaker 3D Printer via Unsplash.