The complaint usually arrives long after the clean: a coating that will not adhere, a lens that hazes under inspection, a bond that passes at build and fails at test. By then the cloth is in a skip and the operator remembers nothing. The delay is the clue, not a reason to stop looking.
Residue is a film, not a stain
A visible mark is easy. The expensive problem is a film a few nanometres thick, spread evenly, invisible under normal light. It changes the surface energy of the part, and that changes how the next process behaves: paint crawls, adhesive lifts, a coating fails a cross-hatch test. The wipe that left it may have looked clean and felt dry.
Films come from three places. The fibre itself releases extractables, especially where a binder or a surfactant was used in manufacture. The solvent carries dissolved material from a previous surface and deposits it as it evaporates. The packaging adds a third source: a plasticiser or an antistatic from the bag, transferred to the sheet before it ever touches the part.
Mechanism one: extractables that migrate
Every non-woven contains something that is not fibre. Spunlace needs energy to entangle, and some constructions use a binder; meltblown and spunbond carry processing aids. These are not defects, but they can migrate into a solvent and then onto a surface.
The test is straightforward: soak a known area of cloth in the process solvent, evaporate a fixed volume, and weigh the residue. Express it as milligrams per square metre, not as a pass or fail. Then compare that number with the residue your process can tolerate. A cloth with a low number is not automatically suitable; a cloth with a high number is not automatically wrong for a utility area.
Mechanism two: re-deposition
A wipe lifts soil and holds it. If the same face returns to the surface, some of that soil goes back. The effect is small per pass and large over a shift, and it shows up as a gradual drift in cleanliness rather than a single failure.
Fold discipline is the control. A quarter-fold gives four clean faces; a six-panel fold gives more. The operator turns the cloth inward between passes and discards it when the faces are used. Training that focuses on the fold, rather than on the product, tends to hold. Training that focuses on the product tends to produce a shelf of unused references.
Mechanism three: condensation and humidity
A cool surface in a warm, humid room collects a water film within minutes. That film can dissolve residues from the air and from the cloth, and then dry slowly, concentrating them. The result is a haze that appears hours or days later, often on the parts that were cleaned first.
Check the dew point of the room against the temperature of the part. If the part is below the dew point at any point in the process, the clean is being undone by physics. This is common in facilities that add a wipe-down step before a part has equilibrated.
How to trace it back
Start with the surface, not the supplier. Take a clean coupon of the same material through the same process, and inspect it at intervals: immediately, one shift later, one week later. If the film appears without the part being touched, the source is either the cloth, the solvent or the air, and the coupon narrows which.
Then run a solvent blank: the same volume of solvent, evaporated on a clean coupon, with no cloth. If the blank is clean, the cloth is implicated. If the blank is not, the solvent or the bottle is the source, and the cloth was never the problem.
How long should a residue test run?
Long enough to cover the delay you observe. If failures appear two weeks after build, a same-day test will miss them. Run the coupon for the full interval before you change the specification.
Is a white cloth safer than a coloured one?
Colour is not a residue indicator. A white cloth can carry the same processing aids as a blue one, and the dye in a coloured cloth is usually bound. Choose on extractables data, not on appearance.
Does pre-wetting remove the problem?
It changes it. The solution is now part of the residue profile, and the pouch adds its own extractables. Pre-wetted formats are more repeatable, not automatically cleaner.
