What an Adhesive Needs in Order to Bond

Why this matters

Most bonds that fail in the field were never bonds. The adhesive was correct for the materials, the tech applied it the way the tube said, and it still let go in three weeks, because one of the conditions a bond actually depends on was missing and nothing on the label mentioned it. That failure comes back as a callback with your name on it, and the second attempt usually gets the same treatment as the first. Knowing what an adhesive physically needs lets you predict the failure before you commit, and lets you read a broken joint and say which condition was absent instead of guessing at a different product.

Before you open the container

Adhesives are chemicals and a fair number of them are aggressive ones. Read Section 8 of the safety data sheet for the exposure controls and gloves the manufacturer specifies for that product, and Section 7 for storage, before you break the seal. Two specifics worth naming because they surprise people:

• Cyanoacrylate bonds skin to skin in seconds and reacts with cotton and wool fast enough to smoke. Wear nitrile, not cotton gloves, and keep it off rags you are holding.
• Acetoxy-cure silicone releases acetic acid as it cures. That vapor attacks copper, brass, and bare electrical contacts. Near either, use a neutral-cure silicone instead, and ventilate the space rather than working with your face over the bead.

Solvent-based adhesives and primers are the other half: they need ventilation and no ignition source in the room, which includes a pilot light on a nearby appliance. Hazard communication for the products you carry runs under 29 CFR 1910.1200, and the SDS is the document that rule requires be available to the tech using the product, not just filed at the shop.

Five conditions, and the bond needs all of them

A bond is not a single property. It is five separate requirements, and the joint is only as good as the one that is missing. The unit of analysis is the individual joint, not the product: the same adhesive on the same two materials can pass in one spot and fail six inches away.

1. Wetting. The liquid adhesive must flow out and make intimate contact with the surface at a molecular scale.
2. A clean, sound surface. Whatever is on top of the substrate is what you are actually bonding to.
3. The cure trigger. Every chemistry needs something specific present to harden, and the joint has to supply it.
4. The right stress mode. The joint geometry decides whether the bondline sees its strong direction or its weak one.
5. Service survival. The cured adhesive has to tolerate the temperature, fluid, and movement it will live in.

Fail any one and the joint fails, regardless of how strong the adhesive is on a data sheet.

Wetting: why the adhesive has to be hungrier than the surface

For a liquid to spread rather than bead, its surface tension has to be lower than the solid's surface energy. That is the whole mechanism. Metals, glass, and ceramics carry very high surface energy and almost anything wets them. Polymers run low, and some run so low that nothing wets them without help.

Those figures are typical published values for untreated surfaces, not a spec for the part in your hand. The practical field test costs nothing: flick clean water on the surface. If it sheets out in a continuous film, you have a high-energy, clean surface. If it breaks into beads, you have either contamination or a low-energy plastic, and an adhesive with surface tension in the same neighborhood as water is going to do exactly what the water did.

For the low-energy group, the fixes are flame treatment, plasma or corona treatment, or a chemical adhesion promoter sold for that plastic. Abrading alone does not raise surface energy. It adds mechanical key, which helps, but a roughened polypropylene surface is still polypropylene.

The cure trigger has to be present in the joint

Each chemistry hardens on a different stimulus, and the joint either supplies it or the adhesive stays liquid in the middle where you cannot see it.

• Cyanoacrylate cures on trace surface moisture and needs a thin bondline. In a dry environment or across a wide gap it skins and stays soft underneath. Acidic surfaces such as bare wood slow it further.
• Anaerobic threadlockers and retaining compounds cure in the absence of air and the presence of active metal ions. On stainless, plated, anodized, or plastic surfaces (the inactive ones) they cure slowly or not at all without the matched primer.
• Moisture-cure polyurethane and RTV silicone pull humidity from the air and cure inward from the outside. In a deep or fully enclosed joint the core can stay uncured for a long time.
• Two-part epoxy and urethane cure on their own mix ratio, so the ratio and the mixing are the trigger. Under-mixed material cures soft in streaks.
• Contact cement needs the solvent to flash off before assembly, and gives you no repositioning after contact.

Temperature moves all of them. As a working rule of thumb, reaction rate roughly doubles per 10 C (18 F) of temperature rise, so a joint made in an unheated garage at 10 C can take on the order of twice as long to reach handling strength as the same joint at 20 C. Manufacturers publish minimum application temperatures for exactly this reason; below theirs, the cure may not complete at all rather than merely running slow.

Shear good, peel bad, and the arithmetic of why

An adhesive bondline is strong in shear and compression and weak in peel and cleavage. This is not a property of a particular product, it is geometry.

Take a bracket foot bonded flat to a panel, 1.0 in wide by 2.0 in long, so 2.0 square inches of bond area. Load it in shear, sliding along the panel, and all 2.0 square inches carry the load together. Now load the same joint in peel by pulling the free end of the bracket away from the panel. The stress no longer spreads. It concentrates in a narrow strip at the leading edge, on the order of 1/16 in wide, which is 0.0625 in by 1.0 in, or 0.0625 square inches. That is 3.1 percent of the bond area carrying essentially the whole load, and once that strip lets go the next strip becomes the new leading edge. Peel unzips.

So the design rule is to convert peel into shear: bond along the direction of load, use a lap rather than a butt, add a mechanical fastener or a lip that takes the prying moment, and where the joint must flex, use a flexible adhesive in a thicker bondline so the strain spreads instead of concentrating.

Reading the break tells you which condition failed

When a bonded joint comes apart, look at both faces before you clean anything. There are only three outcomes and each one names a different problem.

• Adhesive failure. One face is clean bare substrate, all the adhesive is on the other side. The bond never formed. Cause is wetting or contamination, meaning condition one or two.
• Cohesive failure. Adhesive remains on both faces and the split ran through the middle of the glue layer. The bond formed and the adhesive itself was the weak link. Cause is an undersized bond area, the wrong stress mode, an incomplete cure, or service conditions past the adhesive's limit.
• Substrate failure. The part broke and the bond held. The joint was stronger than the material, which is the result you want.

This one look changes what you do next. Adhesive failure means fix the surface and the same product will probably work. Cohesive failure means the product or the joint design is wrong and better prep will not save it.

Two joints, one gate, opposite results

A shop bonds a small metal mounting bracket to the plastic housing of a piece of equipment, twice, on two different units. Same adhesive, same tech, same tube, same afternoon.

Unit A, ABS housing. ABS sits in the moderate band. Water sheets on it after a solvent wipe. The bracket foot is 2.0 square inches and the load path is along the panel, so the joint sees shear. The adhesive is a two-part epoxy mixed by the ratio on the cartridge with a static mixer, applied at 21 C, clamped for the full fixture time. It is still in service.

Unit B, polypropylene housing. Water beads. The tech abrades and wipes anyway and applies the same epoxy. It handles fine, passes the tug test the next morning, and lets go at week three. The break face is clean plastic on one side and a full epoxy shell on the other: adhesive failure, so wetting, not strength. Polypropylene at roughly 30 mN/m never let the epoxy wet it, and abrasion did not change that.

The cost of the difference, in labor hours rather than parts: the return visit ran 1.2 hours including drive time, all of it unbilled warranty effort. Flame-treating the polypropylene or switching to a polyolefin-bonding system would have added roughly 0.3 hour to the original job. That is a 4-to-1 ratio of unbilled recovery effort to the prevention effort, and both sides of that ratio are the same kind of hour, absorbed shop time. It would flip only if the treatment step itself were unreliable, which it is not.

What would change the call: if unit B's bracket had been loaded in shear and the housing had been ABS, the same joint passes. If the housing had been ABS but the bracket cantilevered so the joint saw peel, it fails too, this time cohesively, and no surface treatment fixes it. The material and the geometry are separate gates.

Checking a bond before you leave

Verification is not the tug test, which mostly proves the adhesive skinned.

• Confirm the cure clock, not the feel. Manufacturers publish two different times: handling or fixture strength, and full cure. Full cure is commonly many hours to a day. If the joint will be loaded before full cure, support it mechanically until then and tell the customer the date, not just "tomorrow."
• Prove the mix on a witness bead. With any two-part, squeeze a short bead of the mixed material onto scrap alongside the joint. When the witness bead has cured hard all the way through, the joint core has too. A soft witness bead means the ratio or the mixing was wrong and the joint is suspect no matter how solid it feels on the surface.
• Check the squeeze-out. A continuous fillet of squeeze-out around the perimeter proves you had enough adhesive and enough clamp pressure. No squeeze-out means either a starved bondline or a gap the adhesive never closed.
• Record the substrate, not just the product. Write down which plastic or which coating you bonded to. When a joint comes back, the substrate identity is the first thing you will want and the hardest to recover after the part is off the truck.

References

• 29 CFR 1910.1200, Hazard Communication, for the safety data sheet the product's exposure controls and storage conditions are drawn from
• Adhesive manufacturer technical data sheets for application temperature range, fixture time, full-cure time, and mix ratio; these vary by product and are the only authoritative source for the values
• Trade-standard practice on surface treatment of low-surface-energy polyolefins (flame, plasma, corona, chemical adhesion promoter)
• See related: How to Prepare a Surface So a Bond Holds; Adhesive Selection by Material Reference; Common Construction Adhesives Reference