How do you read an ASTM C1521 sealant field adhesion test?
Summary
An ASTM C1521 field adhesion result is read by failure mode: adhesive failure flags a weak bond, while cohesive or substrate failure points elsewhere. The reading holds for fully cured sealant pulled to its movement-capability strain. ASTM C1521-19(2025) is active as of 23 October 2026, and induced failures do not prove a leak.
ASTM lists C1521-19(2025) as the active edition, last updated 10 October 2025, and BSI lists C1521-19(2020) as withdrawn on 16 October 2025 (ASTM International, 2025; BSI, 2025).
The destructive procedure applies a strain tied to the sealant's maximum published movement capability, or a reasonable strain when that capability is unknown (ASTM C1521-19(2025), 1.2).
ASTM C1521-02, as quoted in one maker's manual, set one test every 100 linear ft in the first 1,000 linear ft of joint, then every 1,000 ft or about once per floor per elevation (Dow Corning, 2007); Tremco's current procedure uses the same rate of one test per 100 ft over the first 1,000 ft (Tremco).
The same manual expects full cure usually within 7 to 21 days before the hand pull test, and longer in cold weather (Dow Corning, 2007).
This reading applies to cured, field-applied weatherproofing sealant joints on buildings. It does not cover structural sealant glazing, which ASTM C1193-25 refers to Guide C1401, or laboratory qualification tests such as ASTM C719-22. Pass criteria for a named sealant come from its maker, not from ASTM C1521.
What does an ASTM C1521 field adhesion test measure on installed sealant?
An ASTM C1521 field adhesion test measures how well an installed sealant bead holds to its substrate compared with the sealant's own cohesive strength, on the actual job, by straining a cut section of the bead by hand. ASTM C1521-19(2025), Standard Practice for Evaluating Adhesion of Installed Weatherproofing Sealant Joints, is the active edition as of 23 October 2026 (ASTM International, 2025). ASTM work item WK98138 is open to revise the non-destructive Technique 2 "for clarity and consistency with Technique 1" (ASTM WK98138), so recheck the edition before writing it into a new specification.
The practice has two procedures. The destructive procedure cuts a tab and pulls it to a strain that "correlates to the strain that the sealant bead may experience when subjected to its maximum published movement capability, when known" (ASTM C1521-19(2025), 1.2). The non-destructive procedure strains the bead in place, and its results "should be either adhesive failure or no failure" (ASTM C1521-19(2025), 1.3). The non-destructive procedure can inspect 100 % of a bead or any area where conditions inconsistent with Guide C1193 are suspected (ASTM C1521-19(2025), 1.4).
The practice states what it leaves out: it "only evaluates the characteristics of the adhesive bond relative to the cohesive strength of the sealant in a particular installation" (ASTM C1521-19(2025), 4.6). It is a bond check. Weathertightness, durability and joint movement are separate questions.
How do you check the sealant is fully cured before an ASTM C1521 test?
Check that an ASTM C1521 test bead is fully cured by confirming the maker's cure time for that sealant, joint depth and weather, then cutting a small section to see that the sealant is cured through its depth. The 2008 edition's scope said: "The sealant to be tested shall be fully cured." (ASTM C1521-08, 1.1). The public summary of the 2019(2025) edition does not repeat that line, so read the purchased text for the current wording.
Cure time is set by the sealant, not the standard. One silicone maker's manual asks for the hand pull test "after the sealant is fully cured (usually within 7 to 21 days.)" and for longer curing before testing after cold-weather installation (Dow Corning, 2007). The same manual gives 7 to 14 days, and up to 28 days or more in cold, dry conditions, for full adhesion of two one-part silicones in structural use (Dow Corning, 2007). Adhesion can lag cure: for a two-part silicone that generally reaches full adhesion within 24 hours, adhesion to a metallic fluorocarbon finish "may take several days" (Dow Corning, 2007).
A practical cure check has four parts. Record the application date, temperature and humidity in the test log. Compare the elapsed days with the product data sheet for that joint depth. Cut through the bead to see whether the core is still soft or tacky. If the core is uncured, reschedule the test; an adhesive failure from an uncured bead says nothing about the installed bond.
How do you read adhesive, cohesive and substrate failure in an ASTM C1521 test?
Read an ASTM C1521 pull by where the break happens: at the bond line, inside the sealant, or inside the substrate, and at what extension it happens. Adhesive failure leaves a clean substrate face, which shows the bond was weaker than the sealant. Cohesive failure tears the sealant body and leaves sealant on the substrate, which shows the bond held at least as well as the sealant. Substrate failure tears the surface material itself, such as a weak coating or friable masonry; the destructive procedure can produce it "where deficient substrate conditions exist" (ASTM C1521-19(2025), 1.2).
The extension at failure matters as much as the mode. Tremco's procedure records "the type of failure that occurs upon full extension to failure and the degree of force required (little or much)" and the percent extension before any failure (Tremco). An adhesive failure at a small fraction of the target strain means a weak bond. An adhesive release only well past the target strain is a different finding, and the maker decides whether it passes.
Pass criteria differ by product. One silicone maker lists, depending on the sealant, a tab pulled to 75 mm (300% extension) without bond loss, a tab pulled to 38 mm (150% extension) without bond loss, or cohesive failure with no adhesion loss (Dow Corning, 2007). The practice itself warns that sealant formulations "may fail in cohesion or adhesion when properly installed", so "the sealant manufacturer should be consulted" (ASTM C1521-19(2025), Note 2).
Reading ASTM C1521 results: result, meaning, action
An ASTM C1521 result turns into an action only when it is read against the strain applied, the cure state and the maker's criterion for that sealant. The table below maps each result to its usual meaning and the first action; the extension criteria in it are one maker's examples, and other sealants carry their own.
Result | Meaning | Action |
|---|---|---|
Adhesive failure well below the movement-capability strain | Weak bond at that location | Stop and find the cause; add tests around it |
Adhesive failure only at or past that strain | Bond near its limit for this sealant | Compare with the maker's criterion, such as 150% or 300% extension (Dow Corning, 2007) |
Cohesive failure in the sealant | Bond at least as strong as the sealant | Record as acceptable if the maker accepts cohesive failure |
Substrate failure | Surface layer weaker than the sealant bond | Assess the coating or masonry, not the sealant |
No failure, non-destructive procedure | No bond defect found at that strain | Record it; recheck if tested in hot weather |
Adhesive failure, non-destructive procedure | Deficient bead at that spot | Repair and widen the inspection |
The key row is the first: an adhesive failure well below the strain that matches the sealant's published movement capability marks a bond that will not carry the joint's design movement. In that case, Tremco advises that work should be stopped to find "the root cause and magnitude of the problem" and that more frequent tests should map the affected area (Tremco). Every destructive test needs immediate repair, since the practice notes it "requires immediate repair of the sealant bead" (ASTM C1521-19(2025), Note 1). A result logged without the cure days and the extension applied cannot be compared with the next round of tests on the same building.
Which installation and surface preparation faults cause ASTM C1521 adhesive failure?
ASTM C1521 adhesive failure at installation usually traces to surface preparation, primer, joint geometry or a bond-breaking layer, but the test alone does not say which. The practice puts it plainly: "If a sealant bead fails adhesively, there is no straightforward procedure for determining the cause." (ASTM C1521-19(2025), 4.4). It lists workmanship, the surface preparation used, the specific sealant, poor installed joint design and poor bond chemistry as possible causes (ASTM C1521-19(2025), 4.4).
Tremco's list of faults that a field adhesion test can detect is more specific: contaminated or expired sealant or primer, improper substrate cleaning, incorrect primer selection, poor priming technique, improper joint configuration, three-sided adhesion from missing or misplaced backing, excessive substrate movement, bond-breaking material at the bond line such as moisture, frost or incompatible coatings, and incomplete joint fill (Tremco). ASTM C1472-16(2022) states that "the single largest factor contributing to non-performance of sealant joints that have been designed for movement is poor workmanship."
Before blaming the product, check the daily log for cleaning method, primer lot, substrate temperature and dew. Look at the failed face: a film or dust layer points to cleaning; a sealant that wets the primer but not the substrate points to the primer step. Compare with the mock-up, which Tremco treats as the reference for good results (Tremco). The same discipline appears in how to show whether a product complaint is really the product: rule out handling, application and conditions with evidence before a product verdict.
Why does an ASTM C1521 failure not prove the sealant joint will leak?
An ASTM C1521 failure does not prove the joint will leak because the failure is induced by the test, at a strain chosen by the tester, on a bead that may never see that strain in service. The practice states it directly: "Since any failures that result from use of these procedures are intentionally induced, they do not necessarily mean that the sealant bead will not perform as a weatherseal." (ASTM C1521-19(2025), 4.6). The reverse also holds. A bead that passes the pull test can still leak through voids, laps, a missing end dam or a three-sided joint that tore in service.
Three conditions decide how much weight a failed pull carries. First, the strain: a failure near the target strain says more than a failure only after a much larger pull. Second, the location: one failure among many cohesive results may be a local cleaning lapse, while a run of failures on one substrate points to a system problem. Third, the substrate pairing: the Walsh Construction study notes that "testing a movement joint with two substrates is very different" from a single-substrate test (Walsh Construction, 2014), so each side of a joint between dissimilar materials should be pulled separately, as one maker's procedure also recommends (Dow Corning, 2007).
Leak claims need water testing or inspection evidence. ASTM C1521 adds the bond side of that picture.
How well does an ASTM C1521 result predict sealant joint performance?
An ASTM C1521 result predicts sealant joint performance only for the bond, at the time of the test and under the temperature of the test day. The practice says the results "can be used to assess the likely performance of the sealant bead and to compare the adhesion of different substrate preparations and sealant materials" (ASTM C1521-19(2025), 4.7), which is a likelihood, not a service life.
Temperature limits the non-destructive reading. The practice states that the non-destructive methods "are least effective during high temperature when the sealant is in a compressed condition" (ASTM C1521-19(2025), 4.8). A no-failure result taken on a hot afternoon on a compressed joint is weaker evidence than the same result taken in mild weather.
Laboratory tests answer different questions. ASTM C719-22 cycles a laboratory specimen through water immersion, movement and temperature, and it "relates only to the performance of the sealant when properly installed with recommended primers" (ASTM C719-22, 5.1). ASTM C920-18(2024) classes sealants by movement capability, from class 12.5 to class 100/50 (ASTM C920-18(2024)), and the maker's published movement capability, often stated as that class, sets the target strain for the destructive pull (ASTM C1521-19(2025), 1.2). In laboratory peel testing, ASTM C794-18(2022) notes that "a cohesive failure is not necessarily better than an adhesive failure, if the adhesive value is sufficient for the application." A field pull that matches the maker's criterion is good evidence for the installed bond; it does not replace qualification data.
What should be checked alongside an ASTM C1521 adhesion test?
Check joint dimensions, backing, fill and the test log alongside an ASTM C1521 adhesion test, because the pull test reads the bond and those items decide whether the joint can move as designed. ASTM C1521-19(2025) lists bead size and configuration, joint movement, workmanship, the adhesive bond and the sealant material as the main parameters of sealant performance (ASTM C1521-19(2025), 4.1). The continuous non-destructive procedure is also "useful in the identification of places of poor joint configuration" (ASTM C1521-19(2025), 4.7).
At each pull location, record:
Joint width and depth against the drawing, with the design basis from ASTM C1472-16(2022).
Backer rod or bond breaker present, and no three-sided adhesion.
Complete fill with no voids, which one maker's procedure asks the tester to inspect at the pull (Dow Corning, 2007).
Sealant and primer lot numbers, application date and cure days.
Substrate, cleaning method and primer, per side for dissimilar substrates.
Failure mode, extension at failure and the maker's criterion it was compared with.
Pre-construction pulls on the actual materials make the field results readable: the Walsh Construction study recommends testing the materials that will be used, since "unannounced manufacturing changes can alter material chemistry and performance" (Walsh Construction, 2014). Lubinpla's performance evaluation task checks first whether a test round was valid, then separates results that installation conditions explain from those that point to the product, and returns one cited document for review.
Key takeaways
Read the failure mode and the extension at failure together.
Test only fully cured sealant, and log the cure days and weather.
Take the pass criterion from the sealant maker for that product.
Trace adhesive failures to cleaning, primer, geometry or backing before blaming the sealant.
Treat a failed pull as a bond finding, not as proof of a leak.