Pull-Off Adhesion Testing for Epoxy Floors

Whether an epoxy floor looks good is visible at a glance, but whether it bonds firmly is not something the naked eye can answer. A coating that looks seamless can still separate from the substrate after a few months of vehicle traffic if the bond between the coating and the concrete was already weak from the start. The pull-off test is a way to check that bond directly, with a pull force figure and an observation of where the failure occurred.

This article presents the general characteristics of the pull-off method as applied to epoxy floors in general. Product-specific technical data for each Kadier system is applied only according to the manufacturers official technical documentation when available; this article does not state product-specific data that has not been confirmed. Standards such as TCVN 9014:2011 and TCVN 12574-2018 are used as general reference bases when assessing materials and application, and do not replace a site survey for each individual project.

1. What does the pull-off test measure?

The principle is very simple: a metal pull-off disc, commonly called a dolly, is bonded firmly to the coating surface with a specialist adhesive. After the adhesive has cured, the test device pulls the disc away perpendicular to the surface, increasing the force gradually until the sample separates. The pull force at separation divided by the disc area gives the bond strength, usually expressed in MPa or N/mm².

The important point is not the single figure alone. The failure location shows which bond is weakest in the whole system: the disc adhesive separating from the coating, the coating separating from the primer, the primer separating from the concrete, or the concrete itself breaking apart. A result with a high figure but failure right at the interface between the coating and the concrete is still a sign that substrate preparation needs to be reviewed, whereas failure within the concrete itself shows that the coating bonds better than the substrate itself.

2. Steps in the procedure

Choosing locations. Test locations should be representative of each different area of use in the project, for example forklift routes, areas exposed to chemicals, and areas that have previously been repaired. Do not test only in the best-looking places. The number of test points and their specific locations should be agreed in advance in the acceptance file, because every test point leaves a mark that must be repaired afterwards.

Preparing the surface and bonding the disc. The surface at the test point is cleaned and lightly roughened so the adhesive bonds well. The disc adhesive must be compatible with the coating and must be given enough curing time according to the adhesive manufacturers instructions. A misaligned disc, insufficient adhesive, or adhesive that has not fully cured will all produce a falsely low result.

Cutting around the disc when needed. For thick coatings, cutting a groove around the disc down to the substrate helps concentrate the pull force on the disc area and prevents the surrounding coating from carrying load as well. For thin coatings, cutting may not be necessary and may even cause edge cracking. Whether to cut or not should follow the test method agreed from the start, so that test points can be compared with each other.

Pulling and recording. The pull device must be placed perpendicular to the surface, with force increased evenly and without jerking. The tester records the failure force, the failure mode, and the proportion of area for each mode, together with photographs of the separated surface. A report that records only the figure, without the failure mode, is missing half the information.

3. How to read the results correctly

There is no single universal pass threshold for all epoxy floors. The pass threshold must be taken from the technical file, contract, or standard that the particular project applies, and must be agreed before testing, not after the results have been seen. When comparing test points with each other, they must be compared under the same conditions: the same type of disc, the same cutting method, and the same adhesive waiting time.

Some common situations: uniformly low results across the whole floor usually point to a system-wide problem, such as a damp substrate, a weak substrate, or an unsuitable primer; low results at only a few points usually point to a local problem, such as contamination at that location during application or a locally weaker substrate. A low result caused by failure of the disc adhesive must be repeated, and must not be counted as a fault of the floor.

It is also necessary to distinguish adhesion testing from other checks. Tap testing only detects areas that have already blistered; it cannot measure the bond strength that remains. Thickness measurement only shows whether the coating is thick enough; it says nothing about the bond. The pull-off test is a locally destructive test, so it supplements, rather than replaces, surface observation and the application file.

4. After testing: repairing the test points

Each test point leaves a pulled-off disc and a damaged surface area. This area must be restored using the correct layer system, from cleaning, priming, and coating, then allowed to cure before being put into use, just like a localized repair location. Skipping this step means creating a new weak point on the very floor that has just been accepted. For details of the patching and recoating procedure for individual areas, see Localized Epoxy Floor Repair: Patching and Recoating Individual Areas, and for how to address the causes of peeling, see Epoxy Peeling, Bubbling and Discoloration: Causes and Fixes.

5. When should the pull-off test be used?

The pull method is most suitable for handover acceptance, for checks before overcoating an old layer, and for verification after repair. Before recoating an old floor, a few pull points show whether the old layer still bonds well enough to be overcoated or must be ground off, instead of deciding by feel alone. During application, early testing on a trial area helps detect substrate problems before the whole floor is coated.

Conversely, destructive testing should not be overused on a floor that is performing well, because every test point needs repair and is a discontinuity in the coating. Periodic checks by observation, tap testing, and monitoring signs of deterioration are usually sufficient for day-to-day operation; the pull-off test is reserved for occasions when quantitative evidence is needed.

To choose a coating system suited to the acceptance requirements of a project, see the RollCoat system for roller-applied floors, the LevelFlow system for self-leveling floors, and the overall acceptance criteria in Epoxy Floor Acceptance: Which Criteria?. For advice on a testing and acceptance plan for a specific project, Request a consultation.

Illustration photo: Los Muertos Crew / Pexels.