A cracked plastic part that was in contact with a fluid presents an immediate ambiguity. The fluid may have chemically degraded the polymer, or it may have acted as a stress-cracking agent without altering the polymer's chemistry at all. Photographs cannot separate these, and the two explanations point at different parties — one at the material and the chemical that attacked it, the other at whatever put the part under sustained tensile stress. The laboratory separates them reliably, and it does so by asking a single question: did the polymer change?
Two mechanisms, one appearance
Chemical attack degrades the polymer directly. Bonds in the backbone are cleaved, or additives are extracted, and the material is measurably different afterwards than it was before. Environmental stress cracking does no such thing. It is a physical process in which a surface-active agent lowers the energy required for a craze to nucleate and grow, allowing fracture at a stress far below the material's rated strength while leaving the polymer chemically essentially intact.
Molecular weight is the primary discriminator
Gel permeation chromatography measures the molecular weight distribution of the polymer. Chain scission from chemical attack shows up as a measurable reduction, typically with a broadened distribution. An ESC failure characteristically shows a distribution indistinguishable from unexposed reference material. This is the single most direct available test, and it depends entirely on having a reference — unexposed material of the same grade and ideally the same lot — because absolute molecular weight varies between grades and production runs.
What the infrared spectrum adds
Spectroscopy detects new chemical species. Oxidation introduces carbonyl absorption that was not there before; hydrolysis of an ester or amide backbone produces its own characteristic changes. In an ESC failure the spectrum of the fracture region should look like the spectrum of the bulk and like the reference. A spectrum showing new functional groups concentrated at the fracture surface is pointing toward chemical involvement rather than a purely physical mechanism.
The fracture surface distinguishes both from overload
ESC produces a characteristic morphology: crazed regions where fibrils have drawn and ruptured, generally with a slow-growth zone and a distinct transition to final fast fracture. Ductile overload shows gross deformation, drawing and shear lips. Brittle mechanical fracture shows a flat surface with radial features pointing back to an origin. These are separable under magnification, and they establish that the failure was ESC rather than simply an overloaded part in the presence of a fluid — which is a third possibility that gets less attention than it deserves.
ESC needs three things and the absence of any one stops it
A susceptible resin, a sustained tensile stress, and a surface-active agent must all be present together. That structure is diagnostically useful: if any leg can be excluded, the mechanism can be excluded. It also means an ESC finding always implicates at least two contributions, which is why these matters rarely resolve to a single responsible party without further work on which leg was the abnormal one.
The agents are usually unremarkable
Aggressive chemicals are not required. Detergents, lubricants, oils, alcohols, adhesives and some cleaning products are documented stress-cracking agents for susceptible resins, and a fluid that is entirely benign against an unstressed specimen can crack the same material readily once stress is present. This is why compatibility data generated on unstressed coupons routinely misleads, and why the stress condition has to be part of any compatibility assessment that is going to mean anything.
Where the stress came from is a separate inquiry
Establishing ESC identifies the mechanism but not the source of the stress leg, which may be residual stress frozen in during moulding, an interference or clamping stress imposed at assembly, or an applied service load. Polarised-light examination visualises residual and assembly stress directly in transparent and translucent materials, and comparing a failed part against an unassembled exemplar often shows how much of the total stress the assembly itself contributed.
Reproduction is the strongest confirmation
Where exemplar parts and the suspect fluid are both available, reproducing the failure under controlled stress and exposure — the bent-strip approach in ASTM D1693 being the common reference — moves the finding from inference to demonstration. A reproduction that generates the same fracture morphology at a comparable stress is considerably more persuasive than a chain of analytical results alone.
What to preserve
The cracked part without cleaning, since cleaning removes the agent that is half the evidence. A sample of every fluid the part contacted, with concentration and exposure conditions if known. Unused parts from the same lot, both for reference measurement and for reproduction testing. And the assembly context — torque values, fits, and how the part was constrained — because the stress leg cannot be reconstructed from a loose part.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.