Read a Coating Adhesion Test Without Overstating the Result

A coating adhesion report can contain a tidy class, rating or stress value while leaving the actual test meaning unclear. Cross-cut and tape methods classify a method-specific response. Pull-off testing reports a stress reached by a test assembly that may fail in the glue, within a coating layer, at an interface or in the substrate. Those outputs cannot be placed on one conversion scale.

This review is limited to coating-test evidence. The existing discussion of damage to brass-coated steel cables inside rubber tracks provides adjacent damage context, but it does not define a paint or coating adhesion test. The actual coating system, substrate, method and acceptance source must come from the order.

Identify the method and result unit

Start with the part and test location, the substrate and its preparation, and the complete coating system. List primer, intermediate and topcoat layers where applicable, along with the specified condition and cure stage. Connect the sample or finished part to the heat, lot, batch or serial trace used by the order.

Transcribe the invoked method and edition. The official page for ISO 2409:2020 describes a cross-cut classification that assesses a coating's resistance to separation from the substrate or another coating. It explicitly does not describe the result as a measurement of adhesion. Its scope also contains applicability limits related to coating thickness and texture that must be checked in the controlled document.

The public page for ASTM D3359-23 covers tape-test methods and their rating systems for stated coating and substrate conditions. A rating is a classification under the selected method. It does not state an absolute bond force and should not be translated into another standard's rating by label similarity.

ASTM D4541-22 addresses method-specific pull-off strength using a loading fixture attached to the coating. Its native result uses force divided by fixture area under the selected protocol and apparatus. That stress value is not a converted cross-cut class, and a cross-cut class cannot be converted to MPa.

Record whether the requirement calls for a classification, a pull-off result, a pass/fail load or another native output. Keep the result symbol and units as reported. Identify whether testing occurred in a laboratory, in the field, on a specimen panel or on the finished part. Record test count, locations and the statistic required by the coating specification.

The acceptance source must be named separately from the method. A test standard explains how to obtain and report an output; the drawing, coating specification, contract or approved plan supplies the limit and disposition authority. This article supplies no universal coating threshold.

Find where the failure occurred

Map the test system before interpreting failure. For a multilayer coating, show the substrate, each coating layer, the glue where used and the dolly or fixture. After a pull-off event, inspect both the dolly face and the remaining test site. The fracture appearance indicates which plane limited the observed result.

Adhesive failure occurs at an interface, such as coating to substrate or between coating layers. Cohesive failure occurs within a layer. Substrate fracture shows that the base material itself participated in the break. These labels should name the actual layers and interfaces rather than use “adhesion failure” for every visible separation.

Glue failure is especially important. If the attachment adhesive fails before the coating/substrate interface, the reported pull value records the level reached by that test assembly. A high value ending in glue failure does not establish the interface strength; the interface may be stronger than the reached value, but the test did not measure its breaking point. Do not replace that limited conclusion with an assumed strength.

DeFelsko's adhesion-testing guidance illustrates adhesive, cohesive and glue-related fracture locations by examining the dolly and surface. It is useful for record structure but does not impose a universal percentage classification or acceptance limit.

Where the report gives mixed failure percentages, retain them exactly as assessed and identify the operator and method. Do not recalculate from an unlabeled photograph. Link photographs to part, location, test number and both damaged faces. If glue obscures the coating or the fracture cannot be classified, say “unclassifiable” or use the procedure's authorized wording rather than selecting the most favorable plane.

For cross-cut or tape testing, record the removed area and native classification under the invoked procedure. In a multilayer system, identify which interface or layer was exposed where possible. The cross-cut observation is not a pull-off fracture map and does not reveal a force-per-area value.

Repeating a test or changing glue, dolly or preparation requires authorization. Preserve the original result and damaged site record. The repeat and disposition owner should decide whether another test is valid and how both results are used.

Check the geometry and test conditions

Describe the test surface as flat, curved or otherwise constrained and record access. Substrate rigidity and local shape can affect cutting, fixture alignment and deformation. A result from a flat witness panel cannot silently represent a finished curved part unless the coating specification establishes that relationship.

Record substrate material and preparation, coating layers, thickness, cure, age and any conditioning. Include temperature and humidity when required by the procedure. The same coating name on a different substrate, preparation or cure state is not necessarily comparable evidence.

For cross-cut or tape tests, retain cut spacing, tool identity and condition, penetration evidence and the controlled method variant. Record tape identity, lot and storage conditions, plus application, removal, peel direction, angle, rate and operator details required by the procedure. ASTM D3359 notes limited resolution and sensitivity to test conditions; the rating should remain within that method-specific scope.

For pull-off testing, identify the ASTM D4541 protocol or applicable method, apparatus, dolly size and shape, glue and cure, scoring or cutting practice, pull rate and alignment. Record equipment calibration or verification status. Different apparatus and fixtures can produce results that should not be compared without the method's required compatibility.

Scoring can change the stress field and fracture path, while omitting it may leave adjacent coating engaged. The report must state what was done rather than allow the reviewer to infer it from the value. Glue selection and cure need enough documentation to explain whether the attachment itself limited the test.

Pull-off and many cross-cut tests damage the tested area. Identify who authorizes the test location and how the site is repaired, protected or excluded from use. Do not conduct an unapproved destructive check on a functional surface. Record every deviation from the invoked procedure.

Record state	Method and native result	Geometry, conditions and failure	Review action
Normal: comparable evidence	Part, coating, substrate, method, edition, protocol and native class, rating or stress output match the requirement.	Location, geometry, preparation, thickness, cure, apparatus or tape variables and actual fracture or removal record are traceable.	Compare only with method-compatible criteria through the authorized coating review.
Missing: result context incomplete	The apparatus, tape, dolly, glue, protocol, test location, count or acceptance source is absent.	Surface condition, cure, scoring, pull settings, removed area, fracture description or linked photographs are missing.	Request the absent evidence or an authorized repeat without inventing the failure plane.
Conflict: incompatible claim	A cross-cut class is converted to pull-off stress, rating systems are translated, or incompatible methods are averaged.	Glue failure is presented as interface strength, or different geometry, coating, apparatus or cure conditions are compared as equivalent.	Preserve native results, state the conflict and route it to the specification and disposition owners.

The normal state means the evidence can enter the specified comparison; it does not release a lot. Missing and conflict states prevent a conclusion until the record or an authorized resolution is available.

Compare only compatible results

Build the comparison set from the same standard, edition, method and protocol. Match substrate, preparation, coating system, layer thicknesses, cure and age. For pull-off results, align apparatus and fixture where the method requires it, along with dolly, glue, scoring, pull rate and environment. For cross-cut and tape results, align cutting and tape variables.

Use the same failure basis. A coating-interface failure, cohesive break and glue-limited result answer different questions even when their displayed pull values are close. Keep individual results and the required statistic visible. Exclude or classify invalid results only through the invoked procedure and preserve their record.

Review repeatability and uncertainty where the report or agreement provides them, especially near an acceptance boundary. Do not invent uncertainty from the spread of a few values or change rounding to obtain conformity. Deviations belong beside the result, not in a detached note.

A 5B-style tape rating and a pull-off result in MPa answer different method-specific questions. Neither is universally better. The coating requirement should select the test that addresses its intended control, and the acceptance owner should apply the corresponding native criteria.

If a report is corrected, retain version history and identify the changed method, location, calculation or fracture assessment. Link every result to the tested specimen or finished part and its lot. A panel result should not be generalized to production parts without the approved coating qualification or sampling relationship.

Location coverage deserves the same restraint. Adhesion can vary near edges, welds, repairs, geometry transitions or areas with different surface preparation. A small number of destructive sites cannot describe every coated surface unless the approved plan establishes that representation. List untested or inaccessible zones and preserve the site-repair record so later inspection does not mistake a repaired test mark for a coating defect.

Adhesion-test evidence also does not prove corrosion life or service durability. Those outcomes involve coating system design, exposure, defects, damage and other specified evaluations. This article does not establish testing capability for any supplier or accept or reject a lot. Its endpoint is a method-and-failure record that shows whether evidence is comparable, incomplete or conflicting without overstating what broke.