The bond test, also called the adhesion test, measures the force needed to separate a coating, paint, mortar, ceramic tile or adhesive from its substrate, and records where the separation happened. The result is given in MPa in the pull-off test, in N/mm in the peel test, or as a visual class in the X-cut and cross-cut tests. For mortar renderings, the standard is ABNT NBR 13528:2019, the Brazilian standard for tensile bond strength of mortar coatings. For paint, it is NBR 11003:2023, and for ceramic tiling, NBR 13753, 13754 and 13755.
This guide covers what the test measures and why the failure mode matters, the types of bond test, how to calculate the results, the procedure and acceptance limits for mortar and ceramic tile, paint and adhesive tests, the standards, and worked examples.
The photo shows a tensile bond test on a mortar rendering. Metal discs are bonded to the wall, and the rendering is cut around each one down to the substrate. One disc has already been pulled off with the mortar disc still attached, and the portable pull-off adhesion tester is fitted to the next one.
What the bond test is and why it matters
ABNT NBR 13528-1 defines bond (adhesion) as the property of a coating to resist stresses at its surface or at the interface with the substrate, and adds a point that is often forgotten: bond strength is not a property of the mortar but of the system (substrate, substrate preparation, coating and finish). The same reasoning applies outside construction: the adhesion of a paint depends on how the steel sheet was prepared, and that of a tape on the surface it is applied to.
When adhesion is insufficient, the coating peels, delaminates or debonds before the end of its expected service life, and the defect usually shows up only after the building is handed over or the product is in use. The test anticipates that problem. It is used for quality control on construction sites and production lines, to compare and select paints, mortars, adhesives and surface preparations, and to develop new products before applying them on a large scale. That is why it appears in civil construction (mortar, ceramic tile, facade painting), in the paint and coatings industry (on metal, concrete, wood and plastic), in the adhesives, tapes and sealants industry, and in the automotive industry, which evaluates paints, adhesives and sealants on components and car bodies.
Where the separation happens matters as much as the force. Failure can be adhesive, at the interface between two layers, or cohesive, within one of the layers (the substrate, mortar, paint or adhesive). Only adhesive failure measures the bond of that interface. In cohesive failure, the interface withstood more than the measured value, and the number obtained is actually the strength of the layer that failed. That is why every bond test result is reported together with the failure mode.
The result also depends on how the test is performed: on surface preparation (cleaning, primer, spatter dash), the age and moisture of the coating, the temperature and the loading rate. And since adhesion varies from point to point on the same wall or part, the standards require several tests per condition and decide on the set, never on a single value.
Types of bond test
The methods differ in how they load the interface and in the type of result they give:
- Tensile bond test (pull-off). A metal piece (the disc, for mortar and ceramic tile; the dolly, for paint) is bonded to the coating and pulled perpendicular to the surface until it detaches. In thick coatings, the area around the disc is cut down to the substrate, so that only the disc of material under it is tested. The result is the maximum force divided by the bonded area, in MPa. On site, the test is performed with a portable pull-off adhesion tester.
- X-cut or cross-cut. Cuts through the paint film down to the substrate create loose edges; an adhesive tape is applied over them and removed, pulling off the poorly bonded pieces. The result is a visual class, based on the detached area, not a number in MPa. It is fast, inexpensive and done on the part itself.
- Peel test. A flexible adherend (tape, film, laminate) bonded to a substrate is folded back at 90° or 180° and peeled at constant speed; in the T-peel test, both adherends are flexible and open into a T. The force concentrates on the line where the adhesive is separating, so the result is expressed per unit width, in N/mm.
- Lap shear test. Two sheets are bonded with an overlap of known area and pulled in the plane of the adhesive, which works in shear. The result is the maximum force divided by the bonded area, in MPa. It is the test for structural adhesives.
The table summarizes the four methods and where each one is used:
| Test | How it works | Result | Where it is used |
|---|---|---|---|
| Tensile bond test (pull-off) | disc bonded and pulled perpendicular to the surface | MPa and failure mode | mortar, ceramic tile, paint, thermal spray coatings |
| X-cut or cross-cut | cuts in the film, adhesive tape applied and removed | visual class | paint |
| Peel test at 90°, 180° or T-peel | flexible adherend peeled at constant speed | N/mm | adhesives, tapes, laminates, labels |
| Lap shear test | bonded joint pulled in the plane of the adhesive | MPa | structural adhesives |
A bond test is not a pull-out test. On Brazilian construction sites, the mortar test is commonly called "arrancamento" (pull-out), but strictly speaking a pull-out test is the extraction of an embedded element, such as an anchor, bar or bolt. That other test is covered in pull-out test.
How to calculate bond strength
In the methods that give a number, the calculation is always a force divided by an area (pull-off and lap shear, in MPa) or by a width (peel, in N/mm). The table brings together the formulas and the areas used in the examples in this article:
| Quantity | Formula | Note |
|---|---|---|
| Tensile bond strength | Ra = F / A | in MPa, to two decimal places (NBR 13528-2) |
| Area of the 50 mm disc | A = π × 25² = 1,963.5 mm² | circular disc of (50 ± 1) mm |
| Paint pull-off (20 mm dolly) | σ = F / 314.16 mm² | ASTM D4541, ISO 4624 |
| Peel strength | P = Fmean / b | mean of the plateau divided by the width, in N/mm |
| Lap shear strength | τ = Fmax / (b·L) | ASTM D1002 |
Mortar bond test: NBR 13528 step by step
ABNT NBR 13528 (the Brazilian standard for tensile bond strength of mortar coatings) has three parts, all from 2019: Part 1 gives the general requirements, Part 2 the bond to the substrate and Part 3 the surface bond; Parts 2 and 3 replaced NBR 13528:2010. The test is the pull-off test described above, performed on the finished coating: the standard sets the age, the number of points, the cutting depth and the equipment, so that results from different sites can be compared. The Part 2 procedure, step by step:
- Age. 28 days for cement-sand or mixed mortars; 56 days for lime-sand mortars.
- Test specimens. 12 per condition tested, at least 50 mm apart and away from corners.
- Cutting. Cut the coating around each disc to 1 to 5 mm into the substrate. Wet cutting gives less variability.
- Disc. Metal, circular, (50 ± 1) mm, bonded at the center of the cut area.
- Pulling. With a force gauge of maximum error 2%, continuous load, centered and perpendicular to the plane. Hook-type couplings are not recommended because they introduce eccentricity.
- Calculation and failure mode. Ra = F / A, to two decimal places, plus the failure mode of each point.
How to interpret the failure mode. In a system without spatter dash, failure can occur in the substrate (A), at the substrate/mortar interface (B), in the mortar (C), at the mortar/adhesive interface (D) or at the adhesive/disc interface (E). With spatter dash, the spatter dash layer and its interfaces are added. This is the distinction between adhesive and cohesive failure applied to the coating: when failure is at an interface, the value is the bond itself; when failure is within a layer (substrate, mortar, spatter dash), the real bond is higher than the value, which is recorded as "Ra > value". When failure is in the adhesive or at the disc, the test never actually loaded the coating: the result is discarded and the point is repeated. The diagram below shows, in cross-section, the layers of the test (substrate, mortar, adhesive and disc) and where each failure mode, A to E, is located.
Bond limits for the render base coat: NBR 13749
NBR 13528 says how to test; the minimum values the render base coat (emboço) must reach are in ABNT NBR 13749:2013, the Brazilian standard for acceptance of mortar coatings on walls and ceilings. The limit changes with the location and with the finish the base coat will receive:
| Location | Finish | Minimum Ra |
|---|---|---|
| Interior wall | paint or base for finish plaster | ≥ 0.20 MPa |
| Interior wall | ceramic tile or laminate | ≥ 0.30 MPa |
| Exterior wall | paint or base for finish plaster | ≥ 0.30 MPa |
| Exterior wall | ceramic tile | ≥ 0.30 MPa |
| Ceiling | ≥ 0.20 MPa |
NBR 13749 acceptance is by count: of 12 tests at an age of 28 days or more, at least 8 must reach the limit. The 1996 edition used 6 tests with at least 4 passing, a rule many texts still repeat.
Bond of ceramic tiling and tile adhesive mortar
What NBR 13753 says. ABNT NBR 13753:1996, a Brazilian standard, covers interior and exterior floor tiling with ceramic tiles laid with cement-based tile adhesive: installation, inspection and acceptance. According to its annex, of six bond strength determinations made at 28 days, at least four must be equal to or greater than 0.3 MPa, at a loading rate of (250 ± 50) N/s. NBR 13754 applies the same to interior walls. For facades, NBR 13755:2017 raised the minimum to 0.5 MPa, with 12 test specimens per 2,000 m²: accepted with at least 8 above 0.5 MPa; between 0.3 and 0.5 MPa, the designer is consulted; rejected with fewer than 8 above 0.3 MPa.
Tile adhesive mortar (ABNT NBR 14081-1:2012). This Brazilian standard classifies cement-based tile adhesives by their bond strength at 28 days:
| Property | AC I | AC II | AC III |
|---|---|---|---|
| Open time | ≥ 15 min | ≥ 20 min | ≥ 20 min |
| Bond strength, normal curing | ≥ 0.5 MPa | ≥ 0.5 MPa | ≥ 1.0 MPa |
| Bond strength, water immersion curing | ≥ 0.5 MPa | ≥ 0.5 MPa | ≥ 1.0 MPa |
| Bond strength, oven curing | not required | ≥ 0.5 MPa | ≥ 1.0 MPa |
The suffix E indicates open time extended by at least 10 minutes, and D, reduced slip. For facades, NBR 13755:2017 requires at least AC III.
Paint adhesion test
For paint, the most common test is the cut test, which assesses adhesion by how much paint the tape pulls off the edges of the cuts. ABNT NBR 11003:2023 (Industrial painting: determination of adhesion by cut methods) is the Brazilian standard for this test, performed coat by coat. It has two methods, chosen by dry film thickness:
- Method A (X-cut), for dry films of 70 µm or more: two 40 mm cuts crossing in the middle, adhesive tape applied and removed. Detachment is rated at the intersection (Y0 to Y4) and along the incisions (X0 to X4), and the result is reported as a pair, for example X1Y2.
- Method B (cross-cut), for films below 70 µm: a 6-blade cutter at 2 mm spacing, forming 25 squares. Rated from Gr0 (no detached area) to Gr4.
- Field sampling: number of tests equal to 1% of the painted area in m² (300 m² requires 3 tests); on pipelines, 1 test every 100 m.
The photo shows a cross-cut on a painted steel sheet, with the multi-blade cutter alongside. The small detachments at the intersections of the cuts are what the classification evaluates: the larger the area of paint removed, the worse the class.
Internationally, ISO 2409:2020 uses a scale from 0 (no detachment) to 5, with cut spacing depending on thickness, up to 250 µm; ASTM D3359-23 uses the 5A to 0A (X-cut) and 5B to 0B (cross-cut) scales. Note: in ISO, 0 is the best result; in ASTM, 5B is the best.
Paint pull-off (ASTM D4541-22 and ISO 4624:2023). A dolly, typically 20 mm, is bonded to the paint and pulled perpendicularly until it detaches, at no more than 1 MPa/s. The result is given in MPa, with the nature of failure as a percentage of area: adhesive, cohesive or in the dolly adhesive. It is the method that gives a number, which is why it has gained ground over the cut tests. On concrete, ASTM D7234 is used. Zinc-rich epoxy primers typically give 12 to 17 MPa (Fragata et al., ABRACO, 2012).
Adhesion of adhesives and tapes: peel test
For tapes, films, labels and laminates, adhesion is measured by peeling: one end of the flexible adherend is clamped in the grip of the universal testing machine and peeled at constant speed, at 90°, 180° or as a T-peel, depending on the standard (ASTM D903, D3330, D1876, D6862 or ISO 8510-2). The result is the mean of the curve plateau divided by the width, in N/mm, and can only be compared between tests at the same angle and with the same adherend. Structural adhesives, which join rigid parts, are not tested this way but by lap shear of an overlapping joint (ASTM D1002, see shear test). Angles, test specimens, calculation, typical values and fixtures are covered in the article on the peel test.
Solved bond test exercises
1. Bond strength of a base coat render. A 50 mm disc failed at 510 N, at the mortar/substrate interface. Ra = 510 / 1,963.5 = 0.26 MPa. Because the failure occurred at the interface, this is the bond strength value. On a painted interior wall (0.20 MPa), the point passes. On an exterior wall (0.30 MPa), it fails.
2. Acceptance under NBR 13749. ABNT NBR 13749 is the Brazilian specification for mortar renderings. This example is a painted facade (limit of 0.30 MPa) with 12 forces in N: 640, 598, 702, 555, 610, 480, 689, 720, 575, 630, 590 and 660. The Ra values are 0.33, 0.30, 0.36, 0.28, 0.31, 0.24, 0.35, 0.37, 0.29, 0.32, 0.30 and 0.34 MPa. Nine values reach 0.30 MPa, and the mean is 0.32 MPa. The lot is accepted because 9 is greater than or equal to 8. The count decides, not the mean.
3. Failure within the mortar. A disc failed at 700 N within the mortar layer: 700 / 1,963.5 = 0.36 MPa. Record Ra > 0.36 MPa, because the interface held more than that. If the failure had been in the adhesive, the point would be discarded and repeated.
4. Pull-off test on a coating. A 20 mm dolly detached at 4.1 kN, with 100% cohesive failure in the zinc-rich epoxy primer: σ = 4,100 / 314.16 = 13.1 MPa. This is within the typical range for that primer.
5. Tape peel test. A 25 mm wide specimen had a peak of 26 N and a plateau averaging 18.0 N over 127 mm: P = 18.0 / 25 = 0.72 N/mm (or 18 N/25 mm). The peak is not included in the average.
Common mistakes in bond testing
- Insufficient core cut. If the cut does not reach the substrate, the surrounding ring of mortar also carries load and raises the value artificially.
- Eccentricity. A tilted disc or a hook-type coupling reduces the result. A study of 530 specimens measured a 25% drop (Costa and Carasek, 2009).
- Changing the disc geometry. In the same study, the 50 mm circular disc gave results 52% higher than the 100 mm square disc. The standard specifies the circular disc.
- Treating a failure in the adhesive as a low result. It is an invalid result, and the point must be repeated.
- Deciding by the mean. NBR 13749 and NBR 13755 accept by count: 8 out of 12.
- Using the old 4-out-of-6 rule for base coat render, or the old 0.3 MPa minimum for ceramic tile facades.
- Testing a damp rendering. Moisture tends to lower the bond strength.
- Mixing up the coating rating scales. In ISO 2409, 0 is the best rating. In ASTM D3359, 5B is the best.
- Using the cross-cut test on a thick film that exceeds the method's thickness limit.
- Comparing peel results from different angles, or expressing peel strength in MPa.
Bond test standards
Which standard applies depends on the system being tested. Mortar and ceramic tile follow the ABNT (Brazilian) rendering standards. Paint follows NBR 11003, or the ASTM D4541 and ISO 4624 pull-off tests when you need a result in MPa. Adhesives and tapes follow the ASTM and ISO peel standards. The table lists the main ones:
| Standard | Application | Edition |
|---|---|---|
| ABNT NBR 13528-1, -2 and -3 | tensile bond strength of mortar renderings (Brazil) | 2019 |
| ABNT NBR 13749 | specification for mortar renderings (limits) | 2013 |
| ABNT NBR 13753 and 13754 | ceramic tile on floors and interior walls | 1996 |
| ABNT NBR 13755 | ceramic tile on facades | 2017 |
| ABNT NBR 14081-1 | cementitious tile adhesive (AC I, AC II, AC III) | 2012 |
| ABNT NBR 11003 | adhesion of industrial coatings by cut methods (X-cut and cross-cut) | 2023 |
| ISO 2409 and ASTM D3359 | cross-cut and X-cut | 2020 and D3359-23 |
| ASTM D4541, ASTM D7234 and ISO 4624 | pull-off adhesion of coatings | D4541-22, D7234-22, 2023 |
| ISO 16276-1 | assessment and acceptance criteria for pull-off adhesion of protective coatings on steel structures | 2025 (replaces the 2007 edition) |
| ASTM C1583 | surface tensile strength of concrete and bond strength of repair and overlay materials, by pull-off | |
| ASTM D903, D3330, D1876, D6862, ISO 8510-2 | peel testing of adhesives and tapes | |
| ASTM C633 | thermal spray coatings |
Bond testing on a universal testing machine
As noted above, on-site mortar testing uses a portable pull-off tester. In the laboratory, the universal testing machine runs bond and peel tests at a controlled speed and records the full curve. The Biopdi line offers a tape adhesion fixture (ASTM D3330), wedge, pneumatic and roller grips, load cells from 0.05 to 1,000 kN, and software that calculates the plateau average per unit width. We evaluate fixtures for laboratory pull-off testing on request, as special projects. The panel below shows the machine, what each part does in a bond test, and the tape adhesion fixture.
The universal testing machine in bond testing
The M2 line includes single-column models from 200 N to 10 kN and dual-column models from 5 kN to 1,000 kN. Each part of the machine does a specific job in the test:
- Moving crosshead peels or pulls at the speed set by the standard
- Load cell small capacity, for just a few newtons
- Fixture tape adhesion, wedge, pneumatic or roller grips
- Displacement force vs. displacement curve
- BioLab software plateau average per unit width
- Control panel and emergency stop position and stop the test
Fixtures for bond testing
The fixtures below mount on the same machine, depending on the specimen and the standard:
Size the load cell to the expected force. Tape peel forces fall between 10 and 50 N. A 50 mm mortar disc at 1 MPa needs about 2 kN, and a 20 mm dolly at 20 MPa needs about 6.3 kN.
| Application | Range | Biopdi model |
|---|---|---|
| Tapes, films, labels (peel) | 200 N to 1 kN | MBIOI-1P-M2 |
| Adhesives, polymers, elastomers | 2 to 10 kN | MBIOI-10-M2 |
| Lap shear on metal, thermal spray coatings | 5 to 50 kN | MBIOII-50-M2 |
Other tests and technical references
Tests performed on the Biopdi universal testing machine
Related technical references
Peel test: 90°, 180°, T-peel and peel strength calculation
The peel test measures the force per unit width needed to separate bonded materials. See 90°, 180° and T-peel, formula, units and ASTM and ISO standards.
Shear test: what it is, how to calculate it, shear modulus and standards
A shear test measures the force that makes two parts of a material slide along a plane. See τ = F/A, single and double shear, modulus G, adhesives, composites and standards.
NBR 7222: splitting tensile test, fct,sp calculation, bearing strips and loading rate in kN/s
NBR 7222:2011 measures the tensile strength of concrete and mortar by diametral compression: fct,sp = 2F/(π·d·L) at 0.05 MPa/s. Strips, calculation, examples and machine.
Frequently asked questions
What is a bond (adhesion) test?
It is a test that measures the force needed to separate a coating, paint, mortar, ceramic tile or adhesive from its substrate. The result is given in MPa (pull-off), in N/mm (peel) or as a visual rating class (X-cut or cross-cut), along with the location of the failure.
What is the mortar bond test?
A 50 mm metal disc is bonded to the rendering, which has been cut down to the substrate. The disc is pulled until failure, and Ra = F / A is calculated per ABNT NBR 13528-2:2019 (the Brazilian standard for tensile bond strength of mortar renderings). The test uses 12 points per condition, at 28 days.
What does NBR 13753 say?
ABNT NBR 13753:1996 is the Brazilian standard for interior or exterior floor coverings with ceramic tiles and cementitious tile adhesive. Its annex requires that, out of six bond strength determinations at 28 days, at least four must be equal to or greater than 0.3 MPa.
What is the minimum bond strength for plaster (render)?
Under ABNT NBR 13749:2013, the base coat render needs 0.20 MPa on painted interior walls and ceilings, and 0.30 MPa on exterior walls or on interior walls finished with ceramic tile. At least 8 out of 12 tests must reach the limit.
What are the types of bond test?
The types are tensile bond (pull-off) with a disc (mortar, ceramic tile, paint), X-cut or cross-cut (paint), 90°, 180° or T-peel (adhesives and tapes), and lap shear (structural adhesives).
Which standard covers paint adhesion testing?
In Brazil, ABNT NBR 11003:2023 applies. It uses the X-cut for films of 70 µm or more and the cross-cut for thinner films. Internationally, ISO 2409 and ASTM D3359 apply. For a result in MPa, use the ASTM D4541 or ISO 4624 pull-off test.
What is the minimum bond strength of AC I, AC II and AC III tile adhesives?
Under ABNT NBR 14081-1:2012, the minimum is 0.5 MPa for AC I and AC II and 1.0 MPa for AC III, at 28 days. AC I does not have to meet the oven-curing requirement.
How do you interpret the failure mode?
A failure at the interface gives the true bond strength. A failure within a layer means the bond strength is higher than the measured value. A failure in the adhesive or the disc means the point is discarded and repeated.
Is a bond test the same as a pull-out test?
On Brazilian job sites, the mortar test is often called a pull-out test ("arrancamento"), but they are different tests. A bond (pull-off) test pulls a disc bonded to a layer. A pull-out test extracts an embedded element, such as an anchor or a bar.
What is the difference between ISO 2409 and ASTM D3359?
ISO 2409 rates from 0 (best) to 5 and removes the tape at an angle close to 60°. ASTM D3359 rates from 5B (best) to 0B, or from 5A to 0A for the X-cut, and removes the tape at 180°. The cut spacing is also different.
How is peel strength calculated?
Divide the average plateau force of the curve by the specimen width: P = F / b, in N/mm. The initial peak is discarded. A 25 mm tape with an 18 N plateau has a peel strength of 0.72 N/mm.