Semiconductor cleanrooms are the least forgiving environment for a consumable. The air is already clean; the concern is what the wipe adds. A particle that would be invisible in a pharmaceutical suite can be a yield event on a 300 mm wafer.
That does not mean every wipe in a fab has to be the most expensive reference available. It means the choice has to be matched to the surface, and the surfaces in a fab are not all equal.
Three surfaces, three answers
Tool exteriors, cleanroom walls and floors, and wafer-adjacent surfaces are different cleaning problems.
- Exteriors and floors: a sealed-edge polyester or a low-lint non-woven is adequate. The goal is particle removal without re-deposition, not trace-metal control.
- Tool interiors and load ports: low-shedding polyester, often pre-wetted, with documented extractables. Swabs reach the edges and vacuum ports a wipe cannot.
- Wafer-adjacent surfaces: wipes are rarely the direct contact tool. Where they are, the specification includes ionic and metallic extractables, and the handling procedure is part of the product.
Most fab wipe consumption sits in the first two categories, which is where cost optimisation belongs. The third category should be specified conservatively and left alone.
Why metals matter more than particles
Particles are visible in the data and relatively easy to control with an edge seal. Metals are harder: sodium, potassium, iron, copper and zinc can be present in fibres, binders, dyes and even the water used in manufacturing.
A wipe that is particle-clean can still be a metal source. If the surface is a gate oxide or a sensitive interconnect, the relevant data is ionic extractables by ion chromatography or ICP-MS, not a particle count.
Ask for the ions and the method. A report that says "low metallic impurities" without naming the species is a claim, not data.
Write the surface into the specification, not just the room class. "ISO 5 wipe" is not a specification; "sealed-edge polyester for load-port exteriors, extractables data on file" is.
Static is a contamination route
Dry wiping generates charge. In a fab, that charge attracts particles to the surface you have just cleaned, and it can damage sensitive devices. ESD-safe wipes and swabs exist for this reason, but the label is not enough.
Check the surface resistivity range and the decay performance, and confirm the measurement method. Then check the handling: an ESD wipe used with an ungrounded operator is a static generator with a nice label.
Swabs do more work than expected
The areas that cause yield loss are often small: a seal face, a connector, a vacuum port, the edge of a load port. A wipe cannot reach them without being folded into a point, which is exactly when edge shedding peaks.
A foam or polyester swab with a defined tip geometry is a better tool for these areas, and the cost per clean is usually lower than the alternative: a fold, a fibre, and an investigation.
| Area | Tool | Key data |
|---|---|---|
| Floors, walls | Sealed-edge polyester wipe | Particle and fibre release |
| Tool exterior | Pre-wetted polyester | Particles, residue |
| Load port, seals | Foam or polyester swab | Particle, ionic extractables |
| Optics, sensors | Ultrasoft wipe or swab | Extractables, abrasion |
| Wafer-adjacent | Specified per process | Metals, ions, particles |
Packaging and transfer
A fab's transfer discipline is usually excellent, which makes packaging a smaller differentiator than in pharmaceutical operations. The exception is single-use items entering a tool: double bagging still matters, and the outer bag should be removable without contacting the inner.
Check that the bag does not shed. A clean wipe in a dirty bag is a contaminated wipe by the time it reaches the tool.
What to ask before the first wipe
Three questions, in order. What surface does it touch? What is the failure mode if it leaves residue — yield, rework, scrap? What evidence do we already hold, and what is missing?
The answers usually produce a two-reference specification: one for general surfaces, one for critical areas, with swabs as a third line item. That is a simpler and cheaper position than specifying everything to the highest grade.
Are cleanroom wipes used directly on wafers?
Rarely. Wafer cleaning is a wet-process step with its own chemistry. Wipes are used on tools, surfaces and packaging, and swabs on small features.
Does a lower particle count mean lower metal contamination?
No. Particle count and ionic/metallic extractables are independent measurements. A wipe can excel at one and be average at the other.
How often should fab wipes be requalified?
On supplier change, on material change, and periodically as part of the quality agreement. Yield data is a useful trigger as well as a calendar.
The fab rule is simple: match the wipe to the surface, keep the evidence at the level the surface demands, and do not let a general-purpose cloth drift into a critical area because it was already open.
