A cohesive fracture tells a more detailed story than most failure reports capture, because the investigation usually stops at “adhesive residue on both sides, cohesive failure confirmed” — the easy part. The harder and more useful question is which of several very different root causes actually produced that fracture, and answering it requires a structured investigation sequence rather than a one-line classification.

Step 1: Document the Fracture Site Before It Gets Disturbed

Before cleaning, cutting, or otherwise handling a failed joint, photograph both fracture faces under consistent lighting and note the fracture’s location relative to the joint’s edges and any geometric features (fasteners, fillet terminations, thickness transitions). Cohesive fractures that initiate at an edge or stress concentration tell a different story than one that appears to have started from an internal void — and once a sample gets handled, cut apart for testing, or cleaned, that spatial evidence is gone for good.

Step 2: Classify the Fracture Surface Morphology

Not all cohesive fractures look the same, and the surface texture itself is diagnostic. A ductile fracture surface — rough, torn, often showing drawn-out fibrils or a matte texture — indicates the adhesive had significant elongation capacity at the moment of failure, consistent with an overload event on an otherwise properly cured material. A brittle fracture surface — smooth, glassy, sometimes showing a fan-shaped or river-pattern texture radiating from an initiation point — indicates the material had lost ductility before it failed, pointing toward embrittlement, under-cure, or operation below the adhesive’s intended flexible-service range. Reporting the approximate percentage of the fracture area that is cohesive versus adhesive versus substrate failure, in the style of the fracture-surface classification approach used in standards like ASTM D5573, gives a quantitative baseline that a single photograph or verbal description doesn’t.

Step 3: Run Instrumental Analysis to Confirm What the Visual Read Suggests

Visual and morphological classification generates a hypothesis; instrumental testing confirms or rejects it. Differential scanning calorimetry (DSC) on a sample from the fracture surface checks whether the cured material actually reached its expected glass transition temperature and cure exotherm profile — a Tg reading meaningfully below the formulation’s specified value is strong evidence of under-cure, distinguishing that root cause from an overload failure on a properly cured material. FTIR spectroscopy, run on the fractured material and compared against a reference spectrum of a correctly cured control sample, can reveal unreacted functional groups (residual epoxide or acrylate peaks) that confirm incomplete cross-linking, or conversely confirm the chemistry matches spec and the failure lies elsewhere. Scanning electron microscopy at higher magnification reveals micro-voids, contamination particles, or filler agglomeration acting as stress concentrators that aren’t visible to the naked eye or under optical microscopy alone.

Step 4: Correlate Fracture Evidence With Process Records

Instrumental data identifies what’s wrong with the material; process records identify why. Pull the cure log for the specific lot or production run involved — oven temperature profile for heat-cure systems, radiometer dose readings for UV systems, or dispensing-ratio verification records for two-part chemistries — and check whether any parameter fell outside the qualified process window around the time the failed part was produced. A DSC result showing under-cure paired with a cure-oven log showing a temperature excursion on the affected shift turns a materials-lab finding into an actionable, correctable process event rather than a vague “the adhesive was bad” conclusion that doesn’t prevent recurrence. Email Us if you want a second opinion interpreting DSC or FTIR results against a specific formulation’s reference data.

Step 5: Distinguish Overload From Degradation Using the Full Evidence Set

A joint can show cohesive failure for reasons that call for entirely different corrective actions. A ductile fracture surface, a DSC result matching spec, and a process log with no excursions, combined with a measured operating stress that exceeded the adhesive’s rated cohesive strength, points toward a design or material-selection issue — the joint needs a higher-strength chemistry or a larger bond area, not a process fix. A brittle fracture surface with a depressed Tg reading and a cure-log excursion points toward a process-control failure that a design change won’t fix on its own. Conflating these two categories — treating every cohesive failure as evidence the material was “too weak” — is a common investigation shortcut that leads to an unnecessary and sometimes counterproductive material change.

Step 6: Verify the Corrective Action With a Repeat Test, Not an Assumption

Once a root cause is identified and a fix implemented — a process correction, a design change, or a material substitution — verify it with the same fracture-classification and instrumental sequence on a new sample built under the corrected conditions, rather than assuming the fix worked because the immediate symptom resolved. A corrected cure profile that still shows a depressed Tg on retest, for example, indicates the process fix didn’t fully address the original under-cure issue and needs further investigation before the corrective action is considered closed.

Building an Institutional Reference From Each Investigation

Logging fracture morphology, instrumental results, and the correlated process record for every cohesive-failure investigation builds a reference library that speeds root-cause identification on the next similar failure — a brittle-surface, depressed-Tg pattern that’s been seen before points investigators toward the process-log check immediately rather than starting the full sequence from scratch. For the underlying distinction between cohesive and adhesive failure modes this investigation process assumes, see our companion guide on understanding cohesive bond failure, and for how thermal-cycling stress specifically factors into a brittle-fracture root cause, see how CTE mismatch causes adhesive bond failure.

Incure supports formal failure investigations with reference DSC and FTIR data for its adhesive systems, so a fracture-surface finding can be checked directly against known-good baseline data rather than an assumption. Contact Our Team to review a specific cohesive-failure investigation with our technical team.

Visit www.incurelab.com for more information.