BlogApplication notes
Why is my plastic bond weak?
- Type
- Application notes
- Published
- 20 November 2026
- Prepared by
- Technical team
- Product group
- Adhesives
Switching to a stronger adhesive is the usual response to a weak plastic bond, but when the cause lies in the surface, the same failure can come back. What actually weakens the bond can be low surface energy, ineffective treatment, surface contamination, migrating plasticiser or crazing from solvent, and two of these can act together.
Sometimes a bonded plastic part on the assembly line comes away with a light pull, and the adhesive sits whole on one face while the other face is bare and smooth. The same adhesive holds well on metal and on other parts, yet fails on one particular plastic or one particular lot. In other cases the joint looks sound after assembly and the call comes weeks later, from product already shipped.
A bond failure rarely stops at one part. A joint that lets go after assembly means reworking the finished product, and one that fails after shipment means returns and lost trust. Blaming the adhesive and simply swapping it can let the same failure return. If the cause lies in the surface being bonded rather than in the adhesive, a new adhesive leaves the cause where it was.
For a bond to form, the liquid adhesive must first wet the surface evenly. Whether a liquid spreads on a solid depends on the solid's surface free energy, the pull its surface exerts on other matter, and on the liquid's surface tension, the pull that holds the liquid together. Classic work concluded that a liquid spreads when its surface tension is below a threshold set by each solid, its critical surface tension (Zisman, 1964). Surface free energy splits into a dispersion part, a weak attraction all molecules share, and a polar part, a stronger attraction between atoms with uneven charge, and strong links such as hydrogen bonds come from the polar side (Owens & Wendt, 1969). Once the surface is wet, adhesion comes from mechanical interlocking, where adhesive sets inside the roughness, from chemical bonding and from thermodynamic attraction at the interface. Many industrial plastics, however, have a low surface free energy and no polar groups at the surface, so they bond poorly by nature (Awaja et al., 2009).
In such cases the technical team sets up the five hypotheses below and rules them out one by one. The first check is always the fracture surface of the failed joint. Record, in the standard's terms, whether it is interfacial failure with adhesive on one face only, cohesive failure with the adhesive layer torn and left on both faces, or substrate failure with the plastic itself broken (ISO, 2022). That record is the first thing that tells whether to look at the surface or at the adhesive.
Hypothesis 1: low surface energy
This hypothesis comes first when the part is a polyolefin, such as polypropylene or polyethylene, or a fluoropolymer. These materials are cheap and resist chemicals, so they turn up in containers and caps, interior parts and pipe fittings. Picture a line that moves an adhesive it used happily on metal straight onto polypropylene mouldings. The adhesive does not spread out but draws in at its edges, and once cured it lifts off in one clean piece under a fingernail.
What happens between the adhesive and this plastic is what happens to water on an oily pan. Polypropylene and polyethylene are chains of carbon and hydrogen only, so their surfaces carry almost no polar atoms such as oxygen or nitrogen. Their surface free energy is low and almost all dispersion (Owens & Wendt, 1969), so a polar adhesive beads up instead of spreading (Zisman, 1964). Less contact area means fewer sites to bond, and the contact that does form rests on weak dispersion forces alone (Awaja et al., 2009). That is why a joint can look sound after assembly yet let go cleanly at the interface under a small load or vibration. Fluoropolymers and silicone rubber have lower surface energy still and are harder to bond (Ebnesajjad & Ebnesajjad, 2014).
The easiest check is the material marking on the part and the drawing. Look for marks such as PP, PE, TPO, which is polypropylene blended with rubber, POM, known as acetal, or PTFE, a fluoropolymer, and compare purchase records to see whether the resin or grade has changed even for the same part. How well the surface wets can be measured by applying test liquids of known surface tension and watching whether the film breaks up, and a standard method exists for polyolefin films. The value alone does not predict adhesion, so read it against experience on the same material (ASTM International, 2023a). A single drop of water that sits up round also serves as a quick before-and-after check (Owens & Wendt, 1969). Repeated clean interfacial failure, with all the adhesive on one face, makes this hypothesis the strong one (ISO, 2022).
If the material turns out to be a polyolefin or fluoropolymer, there are two routes. One is to change the surface: create polar groups by corona, plasma or flame treatment (Awaja et al., 2009; Ebnesajjad & Ebnesajjad, 2014), or apply a thin primer, an adhesion promoter made for low-energy surfaces. Fluoropolymers may need chemical etching (ASTM International, 2025). The other is to move to an adhesive family designed for low-energy surfaces. Whichever route is chosen, run a comparative test such as lap-shear on the real material and confirm the gain in numbers (ASTM International, 2023b).
Hypothesis 2: ineffective treatment
When parts fail even though they are treated, suspect the treatment itself. It shows up with film and parts that arrive corona- or flame-treated, with processes that treat parts by plasma and stack them for a while before bonding, and on lines where treater power or speed changes often. Sometimes plasma-treated parts go into boxes in the store and, when the schedule slips, come out for bonding much later. The treatment and adhesive are the same, yet lots bonded the day after treatment hold and only the lots that sat for a long time fail.
A treated surface does not stay as it is for long. Corona, plasma and flame treatment oxidise chains at the surface and attach polar groups such as hydroxyl and carbonyl (Awaja et al., 2009; Strobel et al., 2003), but polymer chains move a little even at room temperature, turning polar groups inward and pushing non-polar segments out to lower the surface energy. This is called hydrophobic recovery. A long-term study of plasma-treated polypropylene found the largest loss of surface energy and adhesion in the first 30 days after treatment, at a rate that depended on crystallinity (Novák & Florián, 2004).
Too much treatment causes trouble as well. A comparative study found that corona treatment breaks chains as it oxidises them and leaves small, water-soluble oxidised fragments on the surface, while flame treatment oxidises the surface with hardly any such fragments (Strobel et al., 2003). When these fragments build up they form a weak layer; the adhesive bonds to it and can come away with it under load. Too light a treatment leaves too few polar groups, and too heavy a treatment can add to this weak layer (Ebnesajjad & Ebnesajjad, 2014; Strobel et al., 2003).
Start by matching dates. Log the treatment date and bonding date for each lot, and see whether failures come only from lots stored a long time. Measure wetting tension with the same test liquids right after treatment and just before bonding, and compare the two (ASTM International, 2023a). Check the treatment settings against the records: power and line speed for corona, distance and air-to-gas ratio for flame, time and gas type for plasma (Ebnesajjad & Ebnesajjad, 2014). Also look at whether treated faces were stacked against each other, or kept somewhere hot or humid.
Set a maximum time between treatment and bonding; for parts past it, re-measure wetting tension and re-treat if it falls short (ASTM International, 2023a; Novák & Florián, 2004). Fix a range for the treatment settings and log them every shift, keep hands off treated faces, and store parts so treated faces do not touch. If adhesion is low even when parts are bonded straight after treatment, over-treatment may have built a weak layer. In that case, step the treatment level and find the best point with bond tests (ASTM International, 2023b; Strobel et al., 2003).
Hypothesis 3: surface contamination
If failures come only from one mould, one day or one operator on the same material, look at surface contamination. The usual places are injection moulding that uses a release agent so parts leave the mould cleanly, earlier steps that use process oil or rust preventive, and assembly where parts are handled bare-handed. One example is a line where a single mould kept sticking, so operators sprayed release agent on that mould alone. Parts from the other moulds bond well, while parts from that one look just as clean yet let go easily.
Silicone release agents are especially troublesome. Silicone has a very low surface energy (Zisman, 1964), so even a trace film keeps the adhesive from spreading, and the joint parts easily between the film and the plastic. In a study of bonded carbon fibre composite joints, silicone release agent contamination sharply lowered the fracture toughness of the joint; the effect grew with surface silicon concentration, and above a certain level there was no adhesion at all (Markatos et al., 2014). Process oil, rust preventive and skin oil from fingerprints block the adhesive from the plastic in the same way (ASTM International, 2025).
Contamination does not always come from outside. Low molecular weight additives in the plastic, such as slip agents, antistatic agents and processing aids, work their way to the surface over time and form a thin layer (Ebnesajjad & Ebnesajjad, 2014). The layer is invisible, so even a clean part nobody has touched can bond poorly.
A water-break test gives a quick look at oily contamination: run water over the surface, and if the film breaks into patches instead of staying even, oil remains. Stronger evidence comes from splitting one lot in half, cleaning only one half and bond-testing both, which shows the effect of contamination directly in numbers (ASTM International, 2023b). If the problem keeps coming back, ask a test laboratory for XPS, an analysis of the elements in the top few nanometres, to look for silicon or additives (Markatos et al., 2014).
When contamination is the cause, close the way in first. Agree with the moulding side that parts to be bonded will not see silicone release agents, or, if one must be used, fix a paintable grade and the amount. Just before bonding, wipe in one direction with a clean cloth and a solvent that does not attack the plastic, and if needed sand lightly and wipe again (ASTM International, 2025). For materials that bloom additives, shorten storage and clean right before bonding (Ebnesajjad & Ebnesajjad, 2014). Operators wear gloves and keep their hands off the bond face.
Hypothesis 4: plasticiser migration
Treat this hypothesis separately wherever soft vinyl is bonded: flexible PVC sheet or film, or cable sheathing. Its mark is a good initial bond that fails weeks or months later. Consider flexible PVC sheet bonded to a panel and shipped: after a summer its edges lift, and the adhesive in the lifted area feels tacky to the touch.
Flexible PVC holds a lot of plasticiser, an oily low molecular weight substance that makes rigid PVC soft. Plasticiser is not chemically bound to the PVC chains, so it moves over time, leaving by evaporation from the surface, by exudation onto the surface and by extraction into whatever it touches (Chen et al., 2022). An adhesive in contact gives it a bond line to move into. Once there, plasticiser swells and softens the adhesive, lowering its cohesive strength, and where it gathers at the interface it forms a slippery layer between plastic and adhesive.
The problem has long been known. One study reported that plasticiser migration from flexible PVC film is why many pressure-sensitive adhesives cannot be used on vinyl, and that once the migration mechanisms are understood an adhesive can be designed for it (Graziano & Sjöstrand, 1986). How readily a plasticiser migrates is governed by its molecular weight and structure (Czogała et al., 2021). Migration takes time, so the failure shows up late, and it can become marked after a spell in warm conditions.
The clue lies in time. If the initial strength was normal and the bond weakened with time, or over a summer, this hypothesis moves forward. See whether the failed adhesive is tacky, soft or discoloured, and rub the flexible PVC with a finger or paper to see whether oil comes off (Chen et al., 2022). To put a number on it, hold bonded specimens at an elevated temperature for a set time, then test again and compare with the initial value (ASTM International, 2023b).
For flexible PVC, choose an adhesive family made to resist plasticiser, and ask the supplier for plasticiser migration test data (Graziano & Sjöstrand, 1986). If the material can change, consider grades with polymeric plasticisers, whose larger molecules move less (Czogała et al., 2021). If it cannot, try a barrier layer or primer between the PVC and the adhesive. Whatever the route, judge by strength after heat ageing, not by the initial value.
Hypothesis 5: solvent stress crazing
This hypothesis covers the case where the adhesive still holds but the plastic beside it has broken. It occurs where adhesives, primers or cleaning solvents contain solvent, at spots under residual stress from moulding or clamping stress from screws and press fits. Take a housing with screws driven tight into its bosses, where a solvent-based primer went around a boss before a part was bonded on. A few days later, held up to the light, it can show fine cracks spreading around the boss.
Stress and certain chemicals acting together break a plastic far sooner than either one alone. This is environmental stress cracking, and the international standard describes it as sharply shortening the time to failure. The bent strip method in that standard covers cracking caused not only by gases and liquids but also by solids with migrating components, such as adhesives and plasticised materials (ISO, 2006). Where solvent from an adhesive or primer reaches a stressed spot, fine surface cracks called crazes form and grow into cracks.
In this case the adhesive may hold well to the plastic, but the plastic right beside the joint breaks first along the crazes. Plastic fragments come away on the fracture surface, or a web of fine cracks appears around the joint. It may look fine at first and appear late, after repeated temperature changes or loads.
Looking is the first step. Examine the area around the joint, screws and press fits with a magnifier under angled light for fine crazes, and record whether the failure is substrate failure, with the plastic itself broken (ISO, 2022). Crazes clustered where stress concentrates, such as gates, ribs and fastening points, add weight to this hypothesis. Applying the suspect adhesive or primer to bent strips and comparing whether crazes form confirms the cause directly (ISO, 2006).
To stop crazing, first move to a low-solvent or solvent-free adhesive family, and test primers and cleaning solvents on bent strips of that plastic before use (ISO, 2006). Apply only as much adhesive as needed and wipe off squeeze-out at once so it does not sit on the plastic. On the moulding side, review settings or annealing that lower residual stress; on the design side, review shapes that keep clamping stress away from the joint.
Testing the hypotheses
There is no need to investigate all five hypotheses at once. The fracture surface tells the most, so before the failed parts are thrown away, start by recording whether the failure is interfacial, cohesive or substrate (ISO, 2022). If the adhesive tore within its own layer, a cohesive failure, look first at the adhesive rather than the surface: mix ratio, cure time and temperature, and shelf life.
If the failure is interfacial, look at the material first. With an untreated polyolefin or fluoropolymer, hypothesis 1 is the strong candidate. If a treated material still fails at the interface, check wetting tension and time since treatment to test hypothesis 2 (ASTM International, 2023a; Novák & Florián, 2004).
If failures cluster by mould, date or operator, a control test is quickest. Split one lot into cleaned and uncleaned halves and test both; if only the cleaned half improves, it is hypothesis 3 (ASTM International, 2023b; ASTM International, 2025). If the substrate is flexible PVC, the bond failed with time and the adhesive is soft, look at hypothesis 4 (Graziano & Sjöstrand, 1986); if the substrate broke and there are crazes around the joint, confirm hypothesis 5 on bent strips (ISO, 2006).
Once a fix is in, confirm its effect in numbers. After each single change, run lap-shear on the same specimen form and compare with the earlier value (ASTM International, 2023b); if the hypothesis was one that shows only with time, such as plasticiser or a fading treatment, check that the value holds after heat ageing or storage as well.
The hypotheses often overlap. A treated polypropylene part that still carries release agent brings hypotheses 2 and 3 into play together. Change one thing at a time in tests, or there is no telling what worked. Start with cheap, quick steps such as cleaning, and leave larger steps, a new material or adhesive family, until after.
Limits and cautions
Wetting tension targets, treatment settings and adhesive cure conditions differ by material and product, so take the figures from the technical data for the material and adhesive in use. The standard explains that values from small lap specimens are for comparison and should not be used as allowable design stress (ASTM International, 2023b), so validate structural, safety and medical parts on real parts in the real service environment. Bonded joints are strong in shear and weak in peel and cleavage; if the surface is sound and the joint still fails, review load direction and joint shape with a specialist. If failures continue after the cause is narrowed down, or surface analysis or plasma development is needed, work with a test laboratory or an application specialist.
Keep safety in view too. Cleaning solvents, solvent-based adhesives and primers give off vapour, and breathing it can cause headaches, dizziness and nausea, so use them with good ventilation, from open windows or a ventilation system (HSE, 2025). Cyanoacrylate vapour from instant adhesives can irritate the eyes, skin and airways, and a public agency sets an exposure limit for it (NIOSH, n.d.), so wear safety glasses and gloves. Follow the handling and protective equipment sections of the safety data sheet, the SDS, for treatments, primers and adhesives, and leave chemical etching, which uses strong acids and reactive materials, to a facility with the equipment and training (ASTM International, 2025).
Sources
- ASTM International. (2023a). ASTM D2578-23, Standard test method for wetting tension of polyethylene and polypropylene films. https://store.astm.org/d2578-23.html
- ASTM International. (2023b). ASTM D3163-01(2023), Standard test method for determining strength of adhesively bonded rigid plastic lap-shear joints in shear by tension loading. https://store.astm.org/d3163-01r23.html
- ASTM International. (2025). ASTM D2093-03(2025), Standard practice for preparation of surfaces of plastics prior to adhesive bonding. https://store.astm.org/d2093-03r25.html
- Awaja, F., Gilbert, M., Kelly, G., Fox, B., & Pigram, P. J. (2009). Adhesion of polymers. Progress in Polymer Science, 34(9), 948–968. https://doi.org/10.1016/j.progpolymsci.2009.04.007
- Chen, Y., Zhou, S., Pan, S., Zhao, D., Wei, J., Zhao, M., & Fan, H. (2022). Methods for determination of plasticizer migration from polyvinyl chloride synthetic materials: A mini review. Journal of Leather Science and Engineering, 4, 8. https://doi.org/10.1186/s42825-022-00081-8
- Czogała, J., Pankalla, E., & Turczyn, R. (2021). Recent attempts in the design of efficient PVC plasticizers with reduced migration. Materials, 14(4), 844. https://doi.org/10.3390/ma14040844
- Ebnesajjad, S., & Ebnesajjad, C. F. (2014). Surface treatment of materials for adhesive bonding (2nd ed.). William Andrew. https://shop.elsevier.com/books/surface-treatment-of-materials-for-adhesive-bonding/ebnesajjad/978-0-323-26435-8
- Graziano, L. C., & Sjöstrand, S. A. (1986). Development of acrylic latex pressure sensitive adhesive for plasticized PVC films. Journal of Plastic Film & Sheeting, 2, 95–110. https://doi.org/10.1177/875608798600200203
- Health and Safety Executive [HSE]. (2025). SR22, Solvent-based adhesives: Small-scale work. COSHH essentials. https://www.hse.gov.uk/PUBNS/guidance/sr22.pdf
- International Organization for Standardization [ISO]. (2006). ISO 22088-3:2006, Plastics: Determination of resistance to environmental stress cracking: Part 3, Bent strip method. https://www.csagroup.org/store/product/iso_034874
- International Organization for Standardization [ISO]. (2022). ISO 10365:2022, Adhesives: Designation of main failure patterns. https://www.iso.org/standard/80800.html
- Markatos, D. N., Tserpes, K. I., Rau, E., Brune, K., & Pantelakis, Sp. (2014). Degradation of mode-I fracture toughness of CFRP bonded joints due to release agent and moisture pre-bond contamination. The Journal of Adhesion, 90, 156–173. https://publica.fraunhofer.de/handle/publica/236002
- National Institute for Occupational Safety and Health [NIOSH]. (n.d.). NIOSH pocket guide to chemical hazards: Methyl 2-cyanoacrylate. https://www.cdc.gov/niosh/npg/npgd0405.html
- Novák, I., & Florián, Š. (2004). Investigation of long-term hydrophobic recovery of plasma modified polypropylene. Journal of Materials Science, 39(6), 2033–2036. https://doi.org/10.1023/B:JMSC.0000017765.69441.dd
- Owens, D. K., & Wendt, R. C. (1969). Estimation of the surface free energy of polymers. Journal of Applied Polymer Science, 13(8), 1741–1747. https://doi.org/10.1002/app.1969.070130815
- Strobel, M., Jones, V., Lyons, C. S., Ulsh, M., Kushner, M. J., Dorai, R., & Branch, M. C. (2003). A comparison of corona-treated and flame-treated polypropylene films. Plasmas and Polymers, 8(1). https://cpseg.eecs.umich.edu/pub/articles/pap_8_61_2003.pdf
- Zisman, W. A. (1964). Relation of the equilibrium contact angle to liquid and solid constitution. In Contact angle, wettability, and adhesion (Advances in Chemistry, Vol. 43, pp. 1–51). American Chemical Society. https://doi.org/10.1021/ba-1964-0043.ch001