A mid-sized consumer electronics manufacturer in Southeast Asia replaced its isopropyl alcohol wipe step with inline atmospheric plasma cleaning and cut bonding-related scrap from 8.2% to 5.4% in six weeks — a 34% reduction that paid back the equipment in under seven months. The root cause wasn't exotic: their solvent wipe left silicone residues and inconsistent surface energy on polycarbonate housings, and no amount of operator training was going to fix a chemistry problem. Plasma fixed it by activating the surface at the molecular level, right before the adhesive dispenser, every single time.
The manufacturer produced roughly 180,000 units per month of a wearable device — a fitness tracker with a polycarbonate (PC) top housing bonded to an ABS base using a two-part structural acrylic adhesive. On paper, the process was textbook. In practice, between 12,000 and 16,000 units a month failed either the 24-hour shear test on the production floor or returned from distributors with delaminated seams.
Engineering had already tried the usual suspects. They tightened adhesive dispense volume tolerance to ±3%. They validated cure oven temperature uniformity to within 2°C. They even swapped adhesive suppliers twice. Scrap crept down by maybe a point, then crept right back up.
The one variable nobody had instrumented properly was the surface itself. Every housing got a manual IPA wipe about 40 seconds before adhesive dispense. It looked clean. It smelled clean. It wasn't.
The breakthrough came from a two-week surface audit. An engineer pulled 50 housings straight off the line — post-IPA-wipe, pre-adhesive — and ran dyne ink tests plus contact angle measurements on each.
The results were brutal:
IPA was doing what IPA does: it removed loose dust and fingerprints but redistributed low-surface-energy contaminants like silicone oils and mold releases. You can't wipe your way out of a contamination problem when the wipe itself is part of the problem. For a deeper look at why bonding fails on plastics, the Technology & Knowledge library has useful background on surface chemistry fundamentals.
The team evaluated three options: a better solvent system, low-pressure (vacuum) plasma, and atmospheric plasma. They picked atmospheric plasma, and the decision came down to throughput.
Low-pressure plasma does a beautiful job — arguably a more uniform job — but it's a batch process. Load a chamber, pump down, treat, vent, unload. Adding that to a line running 6,000 units per day would have required either a second chamber or a buffer that nobody had floor space for. Atmospheric plasma, by contrast, dropped straight into the existing conveyor with a treatment head mounted 8 mm above the housing surface. Cycle time impact: 2.3 seconds. No pumping. No batching.
For high-mix medical or aerospace work, vacuum plasma often still wins on uniformity and gas chemistry flexibility. For high-volume consumer electronics with consistent geometry, atmospheric is usually the right call. We cover the trade-offs in more detail across our Applications and Capabilities pages.
The physical change was unglamorous. A rotary atmospheric plasma nozzle — roughly the size of a coffee mug — was mounted on an aluminum bracket over the conveyor, 62 cm upstream of the adhesive dispenser. Compressed air (filtered to 0.01 micron) fed the nozzle. A PLC tap from the conveyor encoder gated the plasma on only when a part was present.
Process parameters, after a two-day DOE:
Validation used dyne inks at 56 and 60 dyne/cm, sampled every 500 units. Post-plasma surface energy stabilized at 68–72 dyne/cm — well above the adhesive's wet-out threshold, with variance under 3 dyne/cm unit-to-unit.
Six weeks after go-live, the data told a clean story:
Notice the residual 5.4%. Plasma didn't magically solve every defect — it solved the adhesion-related ones. The remaining scrap came from cosmetic defects, dimensional issues, and electrical test failures. That's important context: plasma is a surface tool, not a cure-all. If your scrap is driven by something other than bonding or coating adhesion, plasma won't touch it.
Three things caught them off-guard — worth noting if you're considering a similar project.
Plasma doesn't literally remove silicone oil layers thicker than a few nanometers. What it does is oxidize and fragment them into polar species that the adhesive can actually bond to. In XPS scans, silicon was still present on the treated surface. The bond strength didn't care.
The nozzle supplier claimed “up to 72 hours” of activation retention. On polycarbonate in their 45% RH environment, they measured meaningful decay within 4 hours and recommended a 30-minute window between plasma and adhesive. For inline integration that was trivial. For anyone planning to plasma-treat in one cell and bond in another, it's a trap.
Not a minor point. The IPA wipe station was a repetitive, fume-heavy task nobody wanted. Removing it improved morale and reduced one ergonomic complaint category entirely.
If you make things that get bonded, sealed, or coated — smartphones, wearables, hearables, small appliances — the pattern here is worth internalizing.
For teams earlier in their evaluation, the Resources section and Whitepapers & Datasheets cover substrate-specific parameter starting points that save weeks of DOE time.
Probably — if your scrap signature looks like theirs. Ask three questions. First, is a meaningful share of your defects adhesion-related (delamination, bond failure, coating lift, ink rub-off)? Second, are you currently using solvent wipes, flame treatment, or corona, and seeing high unit-to-unit variance? Third, do you have 30–60 cm of conveyor real estate near the bonding or coating step?
If you answered yes to at least two, the math usually works. A smartwatch brand we spoke with saw a similar 30% scrap reduction on glass-to-metal bonding. A Bluetooth speaker maker cut gasket seal failures in half after adding plasma before silicone dispensing. The chemistry is the same; only the substrates change.
If you want to sanity-check a specific application against your substrate, adhesive, and volume, the fariplasmatech applications team can run a free feasibility review and, in most cases, a no-cost sample treatment trial. Start a conversation via Contact Us, or browse real substrate-adhesive combinations we've already solved on the Applications page. Scrap you can measure is scrap you can cut — and surface treatment is almost always the cheapest lever to pull.
English
Spanish
Rogatus ad ultimum admissusque in consistorium ambage nulla praegressa inconsiderate
