There is a strange phenomenon on optical module COB production lines: the ionic contamination test exceeds the limit, and production engineers inspect the board three times from every angle—the surface looks perfectly clean, yet they cannot find where the contamination is. The contamination is not on the surface. It is underneath the chip, inside the pad gaps, and in places where wet cleaning liquids simply cannot penetrate.
When it comes to flux residue, what is visible is not the real problem—what is invisible is what matters. In places where liquids cannot reach, the residue sits quietly, waiting for power-up, waiting for burn-in, waiting for three months of operation in a customer’s data center—and then problems begin to appear.
Wet Cleaning Cannot Clean It—But It Is Not the Cleaning Liquid’s Fault
The logic of wet cleaning is simple: deliver the liquid to the contaminated surface, dissolve the contamination, and rinse it away. This logic works perfectly on open surfaces. Flux on solder joint surfaces and large board areas can be addressed by spraying, immersion, and ultrasonics.
However, the structure of COB packaging dictates that the most dangerous residues are precisely not on open surfaces.
When a chip is mounted on a substrate, the gap between the bottom of the chip and the substrate is only tens to one hundred micrometers. The solder mask steps around the pads and the concave corners formed by the dam adhesive are all blind spots for liquids. The surface tension of the cleaning liquid makes it difficult for the liquid to spontaneously penetrate these micro-gaps. Even if it does penetrate, the rinsing water has an even harder time getting in, and the residual cleaning liquid itself becomes a new contamination source. This is why an awkward situation often occurs on production lines: after three cleaning cycles, the surface looks clean, but the ionic contamination test still exceeds the limit.
Plasma Does Not Rely on “Flowing In”—It Relies on “Diffusing In”
The logic of plasma cleaning is completely different from that of wet cleaning. Plasma is an ionized gas. Inside a vacuum chamber, these reactive species—oxygen radicals, argon ions—do not “flow” in a particular direction; instead, they fill every space within the chamber. Gaps under chips, concave corners between pads, and dead zones at the base of the dam structure—wherever gas can reach, plasma can reach.
Once there, what happens? Reactive oxygen radicals in oxygen plasma react chemically with organic residues, decomposing the rosin, organic acids, and activators in the flux into carbon dioxide and water vapor, which are then evacuated by the vacuum pump. Throughout the entire process, there is no liquid involvement, no surface tension barrier, and no “cannot get in” problem.
Can Gas Really Penetrate Such Small Gaps? Here Is the Data
In optical module COB packaging, the gap between the chip and the substrate is typically on the order of tens to one hundred micrometers. Oxygen-containing plasma gas can enter micro-gaps of approximately 100 micrometers between the chip and the substrate and exert a removal effect on organic residues within the gaps.
The thickness of flux residue itself is not large. The flux film remaining after soldering is typically on the nanometer to micrometer scale, and plasma treatment can effectively decompose and remove it within a few minutes. The real difficulty is “not being able to reach it,” not “not being able to remove it.” Plasma precisely solves the problem of “not being able to reach it.”
Plasma Cleaning of Flux Is Not a Cure-All—But It Is Exactly Fit for Purpose
It should be made clear that plasma cannot clean everything. It is effective against organic residues—the rosin, organic acids, and activators in flux are all organic substances, and oxygen plasma decomposes them with high efficiency. However, if the solder joint surface has a relatively thick, already carbonized flux crust, relying solely on plasma will require longer treatment time, and the efficiency will be lower than performing a wet pre-clean followed by plasma treatment.
However, in the COB optical module scenario, the greatest value of plasma lies precisely in this: it is a dry method capable of treating residues inside micro-gaps after packaging.
For COB modules that have already undergone die attach and have formed a dam structure, using plasma as the final cleaning step is the right choice for solving the specific problem of dead-zone residues. It does not need to replace all cleaning steps; it only needs to fill the shortfall that wet cleaning cannot reach.
Remember This in One Sentence
Gas can enter gaps that liquids cannot. Flux is an organic material, and oxygen plasma specifically decomposes organic materials.
Next time the ionic contamination test exceeds the limit and no surface contamination can be found, think about those invisible places underneath the chip and inside the pad gaps. For places that cannot be washed, switch to a method that can “reach” them.
