Reading a Fracture Surface: Adhesive, Cohesive and Substrate Failure

When a bonded assembly comes apart, it leaves behind a complete account of why. The two fracture faces record which interface gave way, what was present at that interface, and how the crack traveled — information that no amount of subsequent process testing can reconstruct. Learning to read that account is the highest-leverage diagnostic skill in bonding and coating work, and it costs nothing but a stereo microscope and the discipline not to clean the parts first.

The Three Modes

Adhesive failure separates cleanly at the interface. One face carries the adhesive or coating; the other is bare substrate. Under magnification the substrate side shows no residue and often retains the texture of the original molded surface. This mode says the bond between the two materials never developed adequate strength — the adhesive itself was never tested. Everything that determines interfacial strength is implicated: surface energy, contamination, treatment adequacy and treatment age, wetting, and the chemical compatibility of adhesive to substrate.

Cohesive failure leaves adhesive on both faces. The crack ran through the adhesive layer rather than along either interface, which means both interfaces were stronger than the bulk adhesive. This is generally the mode you want, because it demonstrates that surface preparation is not limiting the joint. If the joint still fails below its requirement, the corrective action lies in adhesive selection, bond line thickness, cure schedule or joint geometry — not in surface treatment.

Substrate failure pulls polymer away with the adhesive. On a well-designed joint in a tough material this is the best possible result: the bond outperformed the part. On a filled or degraded material it frequently means something else — that a thin, weak layer at the surface of the plastic detached from the bulk beneath it. Glass-filled materials with a resin-rich skin, parts that saw excessive melt temperature, and surfaces carrying a heavy additive bloom all produce this pattern, and all of them are defects rather than successes.

Mixed Mode Is the Normal Case

Real fracture surfaces are rarely pure. A typical failed bond might be sixty per cent adhesive at the edges and forty per cent cohesive in the center, or show adhesive failure over one region of the part and substrate failure over another. The instinct to force a single classification loses most of the value.

Record percentages instead, with photographs at consistent magnification and lighting. The reason is that the trend in the mixture is the most sensitive process indicator available. Increasing treatment power and watching adhesive failure fall from eighty per cent to twenty per cent is unambiguous evidence that treatment was limiting, even if absolute strength has not yet reached specification. A single pass/fail number would have shown nothing.

Spatial distribution carries information too. Adhesive failure concentrated at one end of a joint suggests uneven treatment coverage, an adhesive dispensing pattern that starves that region, or a clamping arrangement that leaves a gap. Adhesive failure around the perimeter with cohesive failure in the center is a classic signature of moisture or environmental attack working inward.

Preserving the Evidence

The single most common way this information is lost is that someone cleans the parts. Solvent-wiping a failed bond removes the contamination that would have identified the cause; pressing the halves back together transfers material between faces and destroys the mode determination; handling the fracture surface adds hand oils that will appear in any subsequent chemical analysis.

The handling protocol worth writing down is short: do not clean, do not reassemble, handle by edges only, bag each half separately in clean packaging, label with part number, date, lot and where in the process or field the failure occurred, and photograph both faces before anything else. Where litigation or a supplier claim is possible, document who held the parts and when — the analysis is only as useful as the confidence that the samples are what they are claimed to be.

From Mode to Corrective Action

The value of classification is that it eliminates most of the possibility space immediately.

• Adhesive failure directs the investigation to surface energy measurement, contamination checks, treatment verification and treat-to-bond interval. Changing to a stronger adhesive is the classic wasted response — a stronger adhesive still cannot attach to an unreceptive surface, and the joint fails at the same interface at a similar load.
• Cohesive failure directs it to adhesive properties and joint design: modulus and elongation against the actual loading, bond line thickness, cure completeness, and whether the geometry is loading the adhesive in peel or cleavage when it should be in shear.
• Substrate failure directs it to the plastic: melt temperature and residence time during molding, additive package and bloom, filler distribution at the surface, and whether the material is the right choice for the stress the joint imposes.

When the Eye Is Not Enough

Visual and low-magnification classification resolves most cases. Where it does not — very thin coatings, transparent adhesives, ambiguous mixed modes — instrumental methods extend the same logic. Infrared spectroscopy identifies what is chemically present on each face, which settles both the mode question and the contamination question at once. Scanning electron microscopy resolves fracture topography and shows whether a crack initiated at a specific defect. Surface analysis quantifies contaminant species at the concentrations that actually matter.

These are worth the cost when the failure is expensive, recurrent, or contested. They are not worth the cost as a substitute for the ten minutes with a stereo microscope that should always come first.

Related Reading

• Adhesive Bond Failures
• Bonding Plastics
• Why Paint and Ink Delaminate From Molded Parts
• Root-Cause Analysis Methods for Plastic Part Failure
• Measuring Adhesion and Abrasion Durability

Frequently Asked Questions