Why Do Molded Parts Turn Brittle — and How Do You Stop It?

Last updated: 2026-09-01 · Reading time: about 10 min

A part that was specified tough and arrives snapping like a biscuit has been changed somewhere between the datasheet and the box — and only three forces make that change: moisture that hydrolyzes the chains, heat history that cuts them, and stress concentrations that give them a place to start.
Brittleness in molded parts is almost never the resin’s fault; it is the drying, the thermal past or the geometry working against it — and each leaves a different fingerprint on the fracture, which is what makes the diagnosis a method instead of a guess. This article walks the fingerprints and the fixes.

Key facts

<>

  • Brittle failure is a fracture appearance, not a material name: the part broke without yielding — smooth, flat, glassy fracture faces — and any ductile resin can be made to fail that way by its processing.
  • Moisture attack is chemical: in polyesters, polyamides, polycarbonate and TPU, water at molding temperatures splits the polymer chains — hydrolysis — and the strength loss is permanent, not a drying artifact.
  • Heat attack is cumulative: excessive melt temperature, long residence, high screw speed and aggressive back pressure each cut chains; the damage adds up across the material’s entire thermal history, regrind included.
  • Stress concentration is mechanical: a sharp inside corner, a weld line across a load path, a thin-thick transition — the material may be perfectly healthy and still crack where the geometry focuses load, as the radius rules explain.
  • The three causes leave different evidence: hydrolysis pairs with drying records, thermal damage pairs with a shifted melt-flow number and discoloration, and mechanical causes pair with where on the part the crack starts.

    • Shrinkage — The dimensional reduction from mold cavity to cooled part. Amorphous resins: 0.4–0.7% (uniform). Semi-crystalline: 1.0–2.5% (directional). Mold makers apply shrinkage factor when machining cavity. Actual shrinkage varies with wall thickness, gate location and packing pressure. See also related guide.

    What does “brittle” actually mean in molding?

    Brittle failure
    Fracture without significant yielding: the part breaks suddenly under impact or bending, with flat, smooth, often glassy fracture surfaces — the opposite of the whitened, stretched, fibrous faces a ductile material leaves.

    That appearance is the first diagnostic fact, because resins are chosen for toughness and molding quietly trades it away. A ductile resin that fails brittle has either changed chemically — its chains cut so they can no longer stretch and pull load along their length — or changed mechanically — its chains locked into an orientation or a frozen-in stress state that turns load into crack energy.
    The distinction matters because the remedies differ: chemical damage is prevented upstream (drying, thermal discipline), while mechanical brittleness is designed out (radii, seam placement, packing profiles). Both are read the same way first: look at where the part broke and what the fracture face shows. A crack that starts at a corner and runs clean tells one story; a fracture across a weld line tells another; a part that crumbles with discoloration tells a third.
    The resin family sets which failure modes are even available — polyesters and polycarbonates can hydrolyze, polyamides can go brittle when bone-dry, and no family is immune to a sharp corner.

    Which cause owns the failure?

    Observation Points to Confirming test
    Smooth fracture + drying records missing or brief Hydrolysis from wet resin. Dry to datasheet spec, run again; compare parts.
    Discoloration (yellow, brown) with brittle faces Thermal degradation — melt too hot or held too long. Compare melt-flow number of system resin against virgin.
    Fracture starts exactly at a weld line Seam weakness on a load path. Relocate or structurally support the seam; see the seam-strength page.
    Crack starts at a corner, boss root or rib junction Stress concentration — geometry, not material. Add radius, review junction mass; retest on new parts.
    Melt-flow number visibly higher than virgin Chain cutting — regrind ratio or thermal history. Run virgin-only shot; check regrind lineage.
    Brittle only in cold conditions Resin grade at its low-temperature limit. Impact testing at service temperatures; grade review.
    PA parts brittle when new, tough after days in air Bone-dry polyamide — dry-state notch sensitivity. Condition or toughen; see the FAQ below.

    The table is a routing map, not a verdict: real failures combine rows, and the combinations tell the story.
    A fracture that starts at a corner in discolored material is thermal damage finding the stress concentration — fix the thermal issue and the corner may still crack, so both rows get worked.
    The order of operations is by reversibility: drying and thermal discipline first, because they stop further damage in every part that follows; geometry second, because it changes where cracks can start; grade selection last, because it is the most expensive change and the least often needed once the first two are clean. What the table cannot replace is the record — drying logs, melt temperatures, regrind percentages.
    Programs that keep them close a brittle-parts case in days; programs that keep them nowhere fund a longer one.

    How do moisture and heat attack the chains?

    1Read the drying record first. Hygroscopic resins — polyamide, polyester, polycarbonate, TPU, PBT — name their drying temperature, time and dew point on the datasheet; the drying page explains why “overnight in the hopper” is not a record.
    Missing or thin records make hydrolysis the leading suspect on principle.
    2Audit the thermal history. Melt temperature against the datasheet window, residence time (shot size versus barrel capacity), screw speed and back pressure — each is a chain-cutting lever, and together they describe the material’s whole day.
    The datasheet work names the window; the machine settings say whether anyone stayed in it.
    3Compare the melt-flow number. A sample of the resin in the system versus the virgin datasheet value: a visibly higher number means shorter chains — degradation or over-worked regrind.
    It is the single most decisive number in the case, and it costs one measurement.
    4Look for the visual witnesses. Discoloration, silver streaks, a burnt smell, splay — the degradation signatures cataloged in the burn-mark page — appear before brittleness gets noticed; their presence upgrades thermal damage from possible to likely.
    5Quarantine and verify. Hold the suspect lot out of production, correct the drying or thermal regime, run verification parts, and retest impact on the new shots — the fix is proven on parts, per the acceptance discipline, before the case is closed.

    How do you design against brittle failure?

  • Radii at every inside corner — sized to the wall share, so no corner becomes the crack starter; the cheapest toughness in the entire design.
  • Seams off load paths — predicted weld lines kept parallel to load or relocated, with structural support where they must stay.
  • Smooth wall transitions — thin-to-thick changes gradual, rib junctions within their mass budget, no hidden thickness cliffs.
  • Dry-state behavior named for PA — polyamide parts specified with their conditioning state in mind, since dry PA is notch-sensitive even when healthy.
  • Service temperature checked at the grade — impact ratings at the coldest the part will see, not the mildest lab condition; filled grades reviewed against the filled-resin margins.
  • Toughness proven on molded parts — drop and impact testing on trial shots, at service temperature, before the tool is signed off.

    • What does a brittle-parts investigation look like?

    It runs as evidence first, opinions later.
    The fracture faces are photographed before anyone touches them — they are the primary documents. The records are pulled in parallel: drying logs, machine settings, regrind percentages, material certificates, and the melt-flow number of the resin in the system versus virgin.
    The routing table above turns that stack into a ranked hypothesis, and the confirming tests run in order of cost: a dried comparison run costs a shift, a melt-flow measurement costs a sample, a geometry change costs a drawing revision. Most cases close inside the first two.
    The discipline that makes it fast is refusing to argue from memory — “we always dry it” is not a drying record — and refusing to change three variables at once, which is how programs pay for the same fix twice. Where the case touches outdoor service, ultraviolet exposure joins the suspect list as its own slow degradation path, covered in the UV page.
    And where the case touches a seam, the repair is structural, not cosmetic: a seam that survives because the load was rerouted is a seam that stays fixed.

    “Brittle parts are usually healthy material with a bad history or a bad corner. The fracture face tells you which — but only if you photograph it before it gets handled to death.”
    — Arthur, Tooling Engineer at aCatMould

    Frequently asked questions

    Can a brittle part be fixed by annealing?

    Only the mechanical kind. Annealing relieves frozen-in stress and can recover toughness where the brittleness came from orientation or packing — worth testing when the fracture face is healthy-looking and the crack starts at a concentration point. It does nothing for hydrolyzed or thermally degraded material: cut chains stay cut, and no thermal treatment reassembles them. That is why the routing table comes first — annealing a hydrolyzed part just bakes the damage in.

    Why do polyamide parts crack in a dry winter and not in humid summer?

    Polyamide draws toughness from absorbed moisture — it plasticizes the amorphous phase and gives the chains room to move. Bone-dry from the dryer or from months in desert storage, the same part is notch-sensitive and impact-brittle, then recovers as it conditions in service air. The design answers: account for the dry state at the impact spec, add radii where the dry part must survive handling, or specify a toughened grade when the duty leaves no conditioning time.
    The part was never damaged — it was just at its driest.

    Is regrind always the culprit?

    Not inherently — controlled regrind at a sane ratio from clean, known material is standard practice, as the regrind page sets out. It becomes a suspect through accumulation: every thermal pass cuts some chains, so high ratios, uncontrolled sources and multiple loops compound into a system resin that has taken more heat history than any datasheet assumed. The melt-flow number is the honest referee — when it drifts from virgin, the blend is degrading, whatever the percentage on paper.

    How do you prove hydrolysis actually happened?

    Three converging tests. The comparison run: dried-to-spec material through the same tool — recovery points to moisture, persistence redirects the case. The melt-flow number: hydrolyzed resin flows faster than its datasheet because its chains are shorter. And the records: a drying log with gaps, a hopper left open, a dew-point meter that was never checked. No single test convicts alone; all three together close the case to an engineering standard.

    Why did parts crack only after months in service?

    Because some cracks need time to grow to visibility. Environmental stress cracking — a stressed surface meeting an aggressive chemical, detergent or fluid — starts invisible and surfaces months later; fatigue at a concentration point counts cycles until it counts out; UV embrittlement accumulates outdoors. A late failure does not acquit the molding: the frozen-in stress or the sharp corner was manufactured in, and service simply collected the debt.
    The investigation is the same — fracture faces, records, geometry — with service environment added as a witness.

    Who owns the fix — resin, process or design?

    Whichever the evidence convicts, in that order of cost. Process owns drying and thermal discipline — the fixes that protect every part without touching anything else. Design owns radii, seam placement and transitions — the fixes that outlive every process drift. Resin owns the remainder: grade changes for genuine service-condition limits, toughened versions where the duty demands them. Programs that buy a grade change before reading the fracture face usually pay twice — once for the material, once for the corner that never got fixed.

    Key takeaways

    1. Brittleness is a fracture appearance with three usual owners: hydrolysis from wet resin, thermal history cutting chains, and stress concentrations focusing load.
    2. The fracture face, the drying record and the melt-flow number close most cases — photograph first, measure second, change last.
    3. Hydrolysis and thermal damage are permanent; annealing helps only mechanically brittle, chemically healthy material.
    4. Design defenses are geometric and free at the desk: radii, seam placement, smooth transitions, dry-state specs for polyamide.
    5. Regrind is a managed input, not a villain — the melt-flow number referees what the ratio cannot.

    If parts are snapping instead of bending, send the fracture face photo and whatever records exist — drying logs, settings, regrind percentages. You will get the cause routed by evidence and the fixes ranked by cost, from a named tooling engineer, before more parts the bin.

    Sources
    MatWeb — material property database, reference for drying specs and impact data
    Plastics Industry Association — industry resources on molding practice
    ISO 20457:2018 — Plastics moulded parts: Tolerances and acceptance conditions, International Organization for Standardization
    Autodesk Moldflow — injection molding simulation, reference for shear and thermal-history analysis

    Sources: Plastics Technology · Plastics Technology Online · Society of Plastics Engineers (SPE) · ThomasNet