Coating Testing-Thickness, adhesion, pinholes, and spark detection

Today, let’s break down four key inspection parameters—thickness, adhesion, pinholes, and  spark detection—identifying the most common errors made on-site and the consequences of those mistakes.
Thickness Inspection: Don't Just Look at the Average

Dry film thickness (DFT) inspection is the first hurdle in the acceptance of anti-corrosion coatings, yet it is also the stage most prone to manipulation.
Let’s start with the standards. ISO 19840 has superseded the traditional ISO 2808, shifting the acceptance criteria from "90/10" to "80/20": 80% of measurement points must meet or exceed the nominal thickness, while the remaining 20% ​​must not fall below 80% of the nominal thickness.
This change is more significant than many realize. While the shift from 90/10 to 80/20 might look like a simple numerical adjustment, it actually raises the bar for compliance. Previously, 10% of points were allowed to fall below 90% of the nominal thickness; now, 20% are allowed to fall below 80%. Although the lower limit dropped from 90% to 80%, the allowance for non-compliant points expanded from 10% to 20%—so, does this represent greater tolerance or a tightening of standards?
In my view, it represents a tightening. Because the 80% lower limit is significantly lower than 90%, any uneven distribution of measurement points makes it more likely that isolated, extremely low values ​​will result in a "fail" verdict. Previously, a reading of 290μm within a 320μm specification might have barely passed; under the 80/20 rule, anything below 256μm is considered a "failed point," and a single failed point can trigger a rework of the entire batch.
The most common mistake made during on-site thickness testing is selecting measurement points primarily on large, flat surfaces while avoiding welds and edges.
Welds on container bottoms, fillet welds on storage tanks, and the backs of gusset plates on steel structures—these areas are difficult to coat during application, so the coating thickness is naturally lower than on large, flat surfaces. If inspection is limited to flat, open areas, the data will inevitably look impressive. But where does corrosion begin? It starts at these thin spots.
There is another hidden issue: uncalibrated thickness gauges. Magnetic thickness gauges produce skewed readings depending on the thickness of the substrate. ISO 19840 requires calibration before measurement using shims of known thickness on a substrate that matches the component's material and curvature. Many inspectors simply use the instrument without ever performing curvature calibration. On the curved surfaces of pipelines and storage tanks, uncalibrated readings can result in errors exceeding 10%.

Adhesion testing: the location of the failure interface matters more than the numerical value.
The pull-off method is the most common technique used in the field for adhesion testing. In accordance with ISO 4624, a test dolly is pulled from the coated surface using hydraulic or mechanical force, and the pull-off strength and mode of failure are recorded.
A significant change in the 2026 regulations is the raising of the initial adhesion requirement from a minimum of 3–4 MPa to a minimum of 5 MPa. This raised threshold has the greatest impact on epoxy primers, as the adhesion of some conventional epoxy zinc-rich primers often hovers between 4 and 5 MPa.
However, the location of the failure interface is even more easily overlooked than the numerical value itself.
The pull-off test records not only the pull-off strength but also a critical piece of information: exactly where the coating failed. Did the failure occur between the coating and the substrate? Between the primer and the intermediate coat? Or within the coating itself?
These three failure modes carry vastly different implications.
Failure between the coating and the substrate indicates inadequate surface preparation, meaning the primer's adhesion is fundamentally flawed. Failure between the primer and the intermediate coat suggests poor control over the recoating interval. Failure within the coating itself points to issues with the coating's cross-link density or cohesive strength.
I encountered this exact scenario during a failure analysis for an offshore wind power project. The measured pull-off strength was 5.2 MPa—a value that appears to meet the standard—but the failure occurred at the interface between the primer and the intermediate coat. This indicates that adhesion between the primer and the substrate was sound; the issue lay in the inter-coat adhesion. A review of application records revealed that the recoat interval for the intermediate coat had exceeded the maximum allowable window. Consequently, effective chemical cross-linking failed to form between the layers, resulting in only mechanical interlocking and insufficient strength.
An adhesion test report that lists only the numerical value without documenting the failure mode is essentially worthless.

Holiday (pinhole) detection: Choosing the wrong method is tantamount to not testing at all.
Pinholes are the most insidious defects in anti-corrosion coatings. Invisible to the naked eye, they allow corrosive agents to reach the substrate directly.
There are two methods for holiday detection, and choosing the incorrect one is more dangerous than performing no test at all.
The wet sponge method is a low-voltage test, typically operating at less than 100V DC. A moistened sponge electrode is moved across the coating surface; when it encounters a pinhole, moisture penetrates the defect to complete a circuit with the conductive substrate, triggering an alarm. This method is suitable for thin coatings (less than 500 μm).
The high-voltage spark method (holiday detector) involves setting the test voltage based on coating thickness. As the probe passes over a pinhole, the high voltage breaks down the air gap within the defect, generating a discharge spark. This method is suitable for coatings thicker than 200 μm.
The most common error in the field is using the wet sponge method on thick coatings. The voltage used in the wet sponge method cannot penetrate coatings thicker than 500 μm, meaning pinholes will go undetected. Conversely, using the high-voltage spark method on thin coatings carries the risk of improper voltage settings puncturing intact coating and creating artificial defects.
Acceptance criteria for pinholes generally allow for no more than three pinholes (with diameters under 0.5 mm) per square meter. However, strictly tracking this figure in practice is difficult; pinholes are typically marked and repaired immediately upon discovery, and few people count the original number of defects after the repairs are complete.
A more practical focus should be on identifying the specific areas where pinholes tend to cluster. If the pinhole density in a specific area is significantly higher than elsewhere, it indicates a systemic issue with the application process—such as an excessive spray distance or high ambient humidity causing premature surface skinning, which traps solvents inside the coating.

Holiday Detection (Spark Testing): Voltage Setting is Critical
While "spark testing" and "pinhole detection" essentially refer to the same process, in an engineering context, "spark testing" usually implies high-voltage holiday detection used for large-scale, continuous inspections, such as on the interior walls of storage tanks or the external anti-corrosion coatings of pipelines.
Voltage setting is the core operational parameter for spark testing.
There are two versions of the commonly used empirical formula in the industry: V = 7843 × √T (where T is coating thickness in mm) and V = 329 × √T (where T is coating thickness in μm). These formulas apply to different scenarios; the former is typically used for thick coatings, while the latter is used for thin coatings.
If the voltage is set too low, tiny pinholes may go undetected, resulting in a "pass" report despite the presence of actual holidays. If the voltage is set too high, it may puncture sound coating, creating a holiday where none previously existed.
Reference figures include a recommended 2.4 kV/mm for epoxy coal tar coatings and 6 kV/mm for polyethylene tape coatings. However, breakdown voltage thresholds vary by product, and ambient application temperature also plays a role—some guidelines suggest reducing the voltage by 5% for every 10°C increase in temperature.
Scanning speed is the factor most easily overlooked on-site. Moving the probe too quickly may result in missing small pinholes, while moving too slowly—keeping the high voltage applied to a single spot for too long—can damage the coating. A steady scanning speed of 0.1 to 0.3 m/s is recommended.
Another operational detail is the necessity of a reliable ground connection. The principle of spark testing relies on completing a circuit through the probe, coating, substrate, and ground. If the grounding is inadequate and the circuit is incomplete, the detector will not trigger an alarm, even though actual holidays may exist. Before testing, it is essential to verify that the grounding resistance meets requirements and that there is no insulation barrier between the component being tested and the grounding terminal.
Two contrasting scenarios
A successful case study.
For an internal anti-corrosion project on a petrochemical storage tank, a comprehensive four-part inspection was conducted during the acceptance phase. Thickness measurements followed the ISO 19840 "80/20 rule," with test points covering large surface areas, welds, tank bottom annular plates, and nozzle roots; each zone included at least five test points. Adhesion testing utilized the pull-off method in every zone, recording pull-off strength and the location of the failure interface; all failures occurred within the coating itself, indicating sound internal cohesion. Holiday (pinhole) detection was performed using the high-voltage spark method, with voltage set according to coating thickness and scanning speed controlled at 0.15 m/s. The inspection report included complete test point distribution maps and raw data.
An inspection conducted three years later revealed the coating to be in excellent condition, with no signs of rust or blistering.
A failed case study.
In another project, the client opted for only thickness and visual inspections to save on testing costs. While the average thickness values ​​in the report looked impressive, the adhesion data was based on the manufacturer's factory tests rather than on-site verification. Holiday and spark testing were skipped entirely.
Two years later, rust spots appeared on the tank wall. Upon cutting into the area, it was discovered that the coating thickness at the source of the rust was only half the design specification, and a pinhole—barely visible to the naked eye—was present. Corrosive media had entered through the pinhole and spread laterally between the coating and the substrate; by the time rust became visible on the surface, the underlying corroded area was already the size of a palm.
Remedial work required stripping and reapplying the coating across the entire tank wall. The initial savings on testing costs were dwarfed by the cost of rework, which was dozens of times higher.
Practical checklist for procurement staff and construction technicians
First, thickness reports must include raw data. Do not accept reports that simply state "Average thickness XXX μm: Pass." Require specific values ​​and location markers for every test point, and clearly indicate the number and percentage of points falling below the minimum thickness limit. Results based on the 80/20 rule must be listed separately.
Second, adhesion reports must specify the location of the failure plane. Simply reporting the MPa value is insufficient; it is essential to clearly indicate whether the failure occurred at the interface or within the coating itself. The specific coating layer involved in an interfacial failure identifies the process step responsible for the issue.
Third, the correct method must be selected for pinhole and holiday (spark) detection. Use the wet sponge method for thin coatings (under 500 μm) and the high-voltage spark method for thick coatings. Record the test voltage and scanning speed in the report, and specify the basis for the voltage setting.
Fourth, welds, edges, and nozzle roots are mandatory inspection points. Coating thickness is naturally lower in these areas, and the probability of pinholes is higher. During acceptance inspections, do not focus solely on large, flat surfaces; identify vulnerable areas and subject them to targeted inspection.
Fifth, sign off on inspection records on the same day. Log all relevant details—ambient temperature and humidity, substrate temperature, inspection time, and the inspector's identity—in the inspection log. Should quality issues arise later requiring traceability, these records serve as the only basis for reconstructing the site conditions.
Ultimately, the four parameters—thickness, adhesion, pinholes, and holiday detection—all serve the same purpose: verifying the integrity of the coating system. Thickness verification confirms the presence of an adequate physical barrier; adhesion verification ensures the coating remains bonded to the substrate; and pinhole/holiday detection checks for defects that penetrate the entire coating thickness.
Skipping any single step breaks the verification chain. A few missing figures in an inspection report today could translate into extra zeros on a rework invoice years down the line.