Why Surface Preparation Decides an Adhesive Joint

Why this matters

A crew bonded twelve mounting pads onto equipment housings over two days. Four came off inside a season. Same adhesive, same lot, same crew, same shop temperature, and the four that released were carrying less load than several that held.

That pattern is the reason this article exists. When a bonded joint fails for a reason that is not the adhesive and not the load, the surface is almost always the answer, and the specific mechanism is not "it was dirty." It is that the adhesive bonded perfectly to a layer that was itself barely attached to the part.

The concept that explains all of it

An adhesive bonds to whatever is on the surface. If a film of oil, mold release, plasticizer that migrated out of a plastic, silicone from a spray, or a loose oxide layer sits between the adhesive and the substrate, the joint's capacity is not the adhesive's strength. It is the strength of that film's grip on the part underneath, which can be almost nothing.

That film is a weak boundary layer, and the entire purpose of surface preparation is to remove it, not to roughen the part. Roughening matters, but it is second: it increases area and gives the adhesive some mechanical key, and it does nothing at all if the contamination is still present when you do it. Abrading a contaminated surface presses that contamination into the fresh material you just exposed, so you finish with a rougher, dirtier surface than you started with.

Before any of this: what preparation itself puts in the air

The preparation step, not the bonding step, is where the exposures are.

• Solvent vapor. Wiping solvents evaporate continuously while you work and while the part flashes off. Ventilate so air moves away from your breathing zone, keep the container closed between passes, and where the safety data sheet calls for respiratory protection, use it under a written program per 29 CFR 1910.134. Read the sheet, which is your right and your employer's duty under 29 CFR 1910.1200.
• Ignition. Most wiping solvents are flammable and their vapor is heavier than air. Do not prepare surfaces near an appliance that can cycle on, an open flame, a hot surface, or work that could throw a spark, and do not perform this work in a pit or a low enclosed space where vapor collects.
• Abrasion. Sanding or grinding a coated part releases the coating. Old coatings on equipment are presumed to contain lead (29 CFR 1910.1025 in general industry, 29 CFR 1926.62 in construction) or hexavalent chromium (29 CFR 1910.1026 and 29 CFR 1926.1126) until assessed, and both are respiratory routes needing respiratory controls, not gloves and glasses. Abrading masonry or concrete releases respirable crystalline silica, covered by 29 CFR 1910.1053 in general industry and 29 CFR 1926.1153 in construction, and wet methods or on-tool extraction are the controls.
• What the instruction itself creates. Telling you to abrade to bare metal means telling you to remove corrosion protection over that footprint. On a part that lives outdoors, in a wet space, or in a corrosive process, that bare footprint is now a corrosion site unless the adhesive fully covers and seals it or you restore the coating at the edges. Decide that before the sandpaper comes out.
• Where the part is electrical equipment, de-energize and verify absence of voltage under 29 CFR 1910.333(b)(2), proving the tester live-dead-live per NFPA 70E-2021, 120.5, which binds through your employer's electrical safety program rather than on its own. Do not do wet testing of any kind on or above energized equipment.

The order, and why it is that order

1. Clean first, before any abrasive touches it. Remove bulk contamination with a solvent appropriate to the substrate and the contaminant. This is the step that keeps step 2 from making things worse.

2. Abrade. Fresh surface, more area, and on many metals a break-up of the aged oxide layer. Use abrasive media that has not been used on other materials, and avoid papers with a stearate or other lubricant loading, because that loading is a release agent you are applying by hand.

3. Clean again. Abrasion debris is loose particulate, and loose particulate is a weak boundary layer with a different name.

4. Bond within the window. Freshly prepared metal starts re-oxidizing immediately, and aluminum in particular re-grows an oxide quickly. The manufacturer states a window between preparation and bonding, and where the job cannot meet it, an adhesion primer applied right after preparation is what holds the surface until you can.

The wipe, done the way that actually removes anything

Most techs wipe. Fewer wipe in a way that takes the contamination off the part rather than moving it around.

• Two cloths. Wet the first with clean solvent, wipe in one direction, then dry with a second, clean, dry cloth before the solvent evaporates. The solvent dissolves the contaminant; if you let it evaporate on the part, it deposits that contaminant right back onto the surface, spread more evenly than before. A wipe that is allowed to air dry is often worse than no wipe.
• One direction, then a fresh face. Wiping back and forth returns what you just picked up.
• Never dip a used cloth back into the container. That is how one contaminated part contaminates every job in the van for the rest of the month, and it is invisible.
• Solvent onto the cloth, not onto the part, so runoff does not carry contamination into a seam or under a lip where it will wick back out.

Silicone deserves its own line. Sprays, some lubricants, some sealants, and some hand products transfer silicone in quantities far too small to see, it spreads readily across a shop, and ordinary wiping is poor at removing it. If bonded joints in your shop have started failing for no reason, look for what silicone-bearing product entered the building.

Where removal is not enough

Some plastics are not contaminated, they are simply low-energy: the adhesive cannot wet them no matter how clean they get. A sibling article covers the wetting condition itself. On the preparation side the answer is a treatment that changes the surface chemistry rather than cleaning it: an adhesion promoter or primer formulated for that polymer family, or a physical treatment where the equipment and training exist. Sanding a low-energy plastic makes a rough low-energy plastic.

The case, followed through

Back to the twelve pads. Four failures, eight holding, so a 33 percent failure rate on a job that had never given trouble before.

The adhesive lot was eliminated first, because the same container did all twelve and the failures were not consecutive.

Cure conditions were eliminated next: same shop, same two days, and the failures split across both days.

Load pointed the wrong way. Three of the four failures were on the lightest-loaded pads, which is evidence against a strength problem and toward an interface problem, because an interfacial failure does not need much load once it starts to release at an edge.

The failure faces settled it. All four came off with the cured adhesive fully intact and a clean, unmarked footprint left on the housing. Adhesive on one side, undisturbed substrate on the other, is the signature of a joint that never bonded to the part in the first place. The eight that held had, on a sample pried off deliberately, adhesive left on both faces.

What actually differed: the four failures were all prepared in the last hour of the second day, using a wiping cloth that had been in service since morning and re-dipped into the solvent container. Everything else in the process was identical.

The controlled retest, on twelve coupons cut from the same housing material:

Group B carried about 40 percent of what Group A carried, and the failure mode is a stronger finding than the load number: 6 of 6 interfacial failures means the adhesive never had a grip on the part, and the loads those coupons did reach were being carried by contamination. Group A's single interfacial coupon is worth noting rather than smoothing over; it was the last one prepared, which is a small hint in the same direction and not a result on its own.

The corrective action was not a different adhesive. It was three lines added to the work instruction: fresh cloth every pass, dry before the solvent flashes, and no cloth returns to the container.

How to verify you got this right

• The water-break test. On a prepared metal surface where water is compatible with the part and cannot enter equipment or reach anything electrical, flood a small area with clean water. A continuous unbroken film means the surface is clean and high-energy. Water that beads or retreats into islands means there is still a low-energy film present. Dry the surface completely before bonding, and where the part cannot be wetted, run the test on a coupon prepared the same way instead of on the part.
• Read the break on every failed bond, before you change products. Adhesive on both faces is a cohesive failure and a genuine strength or geometry question. A clean substrate footprint is a preparation failure, and changing adhesives will change nothing.
• Check the consumables, not just the procedure. Which solvent, which container, how old, which abrasive, how many parts since the cloth was changed. Those are the variables that move between a joint that holds and one that does not, and none of them appear on a work order.
• Prepare a coupon alongside the job from the same material with the same hands and the same cloths, and pull it. It converts an argument into a measurement.

References

• 29 CFR 1910.1200, hazard communication, for the safety data sheet of the specific solvent and adhesive
• 29 CFR 1910.134 for respiratory protection under a written program where the safety data sheet requires it
• 29 CFR 1910.1025 and 29 CFR 1926.62 (lead), 29 CFR 1910.1026 and 29 CFR 1926.1126 (hexavalent chromium), 29 CFR 1910.1053 and 29 CFR 1926.1153 (respirable crystalline silica), for abrasive preparation of coated or masonry surfaces
• Adhesive manufacturer documentation for the required surface preparation, the time window between preparation and bonding, and any specified primer
• See related: What an Adhesive Joint Depends On; What an Adhesive Needs in Order to Bond; Adhesive Selection by Material Reference