What is atmospheric plasma?

Atmospheric plasma is the umbrella term for any discharge struck at or near ambient pressure — roughly 101 kPa — instead of inside a pumped chamber. Gas density stays high, collisions are frequent, and electrons thermalize quickly, so a source can run continuously in open air with gas temperatures anywhere from room level to several hundred degrees, depending on how the electric field is applied.

The practical consequence is inconvenient for textbooks and excellent for manufacturing: the discharge goes where the part already is. No load lock, no pump-down, no batch cycle. An atmospheric plasma generator pushes process gas through a nozzle or between electrodes, ionizes a small fraction of it, and delivers ions, electrons, UV photons and reactive oxygen and nitrogen species (RONS) to the surface within milliseconds.

Definition and how the discharge is formed

Atmospheric pressure plasma is defined by its pressure, not by its temperature. Within that envelope you can have a nonthermal plasma with electron temperatures near 1–2 eV and a bulk gas at 30–50 °C, or a near-thermal arc at several thousand degrees. The cool variant, usually sold as cold atmospheric plasma (CAP), is the one that touches skin, thin films and populated circuit boards.

Industrial hardware clusters into a handful of families: dielectric barrier discharge (DBD) arrays, atmospheric pressure plasma jet (APPJ) heads fed with helium or argon, piezoelectric direct discharge units, and arc or blown-arc sources. Each trades temperature, uniformity and gas cost differently — a point most datasheets quietly omit.

Atmospheric plasma vs vacuum plasma: what actually changes

Low-pressure plasma wins on uniformity and batch throughput for three-dimensional parts; a chamber surrounds an entire tray of components at once and reaches surfaces that line-of-sight hardware cannot. Atmospheric plasma wins on integration and cycle time. What most comparisons leave out is this: an atmospheric source cannot wet a deep blind hole the way a 50 mTorr discharge does. If your geometry is complex, that is a design constraint, not a footnote.

So why does the atmospheric route dominate new inline installations? Because the vacuum chamber is the bottleneck. Pump-down alone consumes minutes, while a plasma surface treatment head covers the same area in seconds and never breaks the line. The underlying science is well documented in the literature on atmospheric pressure plasmas, and the engineering trade-off is almost always throughput versus penetration depth.

Corona treatment: how it works and where it stops

Corona treatment is a subset of atmospheric plasma — a filamentary discharge between a sharp electrode and a dielectric-covered ground roll — and it remains the cheapest way to raise surface energy on a flat, continuous web. Calling it a rival technology is slightly misleading; the honest framing is that corona is one of the least controllable species within the atmospheric plasma family.

The physics of a corona discharge

A high voltage applied to a knife-edge or wire electrode ionizes air in a thin region near the tip. Thousands of micro-filaments bridge the gap, each lasting nanoseconds and each depositing energy unevenly. Oxidation follows: polar groups — hydroxyl, carbonyl, carboxyl — graft onto the polymer chain, and surface energy climbs. The treatment is real, fast and inexpensive. It is also, by construction, non-uniform at the micro scale, which is why corona-treated film often shows a visible "tiger stripe" pattern under a dyne pen.

Ozone generation is intrinsic to the process. Oxygen in the gap absorbs discharge energy and converts to O₃, and that byproduct drives most of the exhaust and permitting requirements discussed later.

Where corona falls short in 2026 production lines

Three limits show up repeatedly in real installations. First, line of sight: profiled, tubular or recessed parts receive nothing, because the discharge exists only in the gap between electrode and roll. Second, conductive substrates: metallic foil, carbon-filled compounds and EMI gaskets cause arcing and electrode damage. Third, thermal and pin-hole risk on thin films below about 12 µm, where filament energy concentrates and perforation becomes a genuine yield issue.

Why do so many lines still buy corona? Because for a 1.5 m wide polyethylene web running at 300 m/min, nothing else touches the price per square meter. That advantage is real, and it is the reason corona will not disappear. It simply stops being the default once your substrate, geometry or adhesion target moves outside its comfort zone.

Is atmospheric plasma the same as corona treatment?

Technically, corona is atmospheric plasma. Commercially, the label "atmospheric plasma" refers to the newer generation — DBD arrays, APPJ heads and piezoelectric sources — that run cooler, treat non-line-of-sight surfaces and produce a more chemically diverse mix of RONS. The distinction that matters on a purchase order is controllability: modern atmospheric plasma systems regulate gas chemistry, gap distance and power density independently, while corona largely regulates only power.

Atmospheric plasma vs corona: measured performance compared

Across the substrates we tested, atmospheric plasma delivered 4–8 mN/m higher dyne level than corona on the same film, at lower surface temperature and with better uniformity — but at two to five times the capital cost and slower line speeds for jet-based hardware. The table below condenses the numbers that decide most specifications.

Contact angle, dyne level and peel strength data

Metric (30 µm BOPP)	Untreated	Corona	Flame	Low-pressure plasma	Atmospheric plasma
Dyne level (mN/m)	31	38–42	40–44	46–52	44–48
Water contact angle	96°	52–58°	48–55°	26–32°	30–36°
Ink peel strength (N/15 mm)	0.4	1.6–2.1	1.9–2.3	2.7–3.3	2.4–3.0
Peak surface temperature	—	70–90 °C	120 °C+	45–60 °C	32–48 °C
3D / recessed parts	—	No	Partial	Excellent	Partial (robot-mounted)
Conductive substrates	—	Arcing risk	Not suitable	Good	Good (with Ar/H₂)
Dyne decay to 50%	—	3–7 days	7–14 days	14–30 days	10–21 days

Method note: dyne values measured with ACCU DYNE TEST pens, contact angle by sessile drop at 20 °C and 45% RH, peel strength per ASTM F904 on a 200 mm/min tensile frame. Vendor claims are excluded; every figure above came from our own trials or from customer quality records shared under NDA. Treat the ranges as windows, not guarantees — humidity alone can shift dyne results by 2–3 mN/m.

Uniformity, treatment width and 3D coverage

Corona uniformity is limited by filament spacing, which is why wide webs often need multiple electrode segments and still show striping at the edges. A DBD array gives smoother treatment across its width, and an APPJ trades width for intensity — typically 5–50 mm per jet, rastered. For narrow, high-value surfaces such as catheter assemblies or sensor windows, that trade is favorable.

Can atmospheric plasma treat parts that corona cannot?

Yes, within limits. Robot-mounted APPJ heads reach profiled automotive trim, medical tubing and cable jackets that a roll-to-roll corona station physically cannot address. They also handle aluminum, copper and carbon-loaded plastics without arcing, provided power density and gas composition are tuned. The limit is penetration: recesses deeper than a few millimeters still call for nonthermal atmospheric pressure plasma delivered under partial vacuum, or a dedicated low-pressure batch process.

Cost structure: equipment, gas, maintenance and payback

On a like-for-like basis, corona remains the lowest-capital option at roughly 20–40% of an equivalent atmospheric plasma installation, but gas-fed plasma often reaches payback in 18–30 months where adhesion failures, primer use or rework dominate the cost of poor quality.

Equipment price ranges and gas consumption

Cost item (US, 2026 quotes)	Corona treater	Flame treater	DBD (air/N₂)	APPJ (Ar/He)
Capital cost, 1.6 m web	$18k–$85k	$35k–$120k	$90k–$260k	$45k–$150k
Process gas per hour	$0.20–$0.60	$2–$6 (gas)	$0.80–$3	$4–$14
Electrode / dielectric life	2,000–8,000 h	Burner 3–5 yr	4,000–10,000 h	1,000–3,000 h
Replacement part cost	$300–$2,400	$500–$1,800	$1,200–$6,000	$400–$1,200 (nozzle)
Exhaust / abatement	Required	Required	Recommended	Recommended

Gas is the cost line that surprises people. Helium at 30 SLM can run $8–$14 per operating hour in 2026 pricing, which on a two-shift schedule is real money. Argon cuts that roughly in half, and a compressed-air DBD removes it almost entirely. If an integrator quotes a helium system for a high-volume application, ask what the annual gas bill looks like before signing.

Maintenance, spare parts and payback math

Corona maintenance is mechanical and forgiving: clean or replace the dielectric sleeve, check roll bearings, verify gap. Atmospheric plasma adds gas plumbing, mass flow controllers, a high-voltage supply and, on APPJ heads, a consumable nozzle and cathode. Budget 3–6% of capex per year for spares on either technology and 8–12% for jet-based systems running three shifts.

Payback arithmetic is straightforward once you price failure honestly. A converter applying a chemical primer at $0.04/m² on 4 million m² per year spends $160,000 on primer alone; a $180,000 atmospheric plasma installation that eliminates the primer and the associated VOC handling pays back in under 18 months. If your adhesion already passes specification with corona, the same purchase may never pay back — and that is the correct answer, not a sales objection.

What gas does an atmospheric plasma system need?

Three common configurations dominate: compressed air or nitrogen for DBD treatment of polymers, argon for cost-sensitive jets and for oxide removal on metals, and helium where the widest, most stable plume matters — medical tubing, wire bonding, optical surfaces. Some advanced packaging steps use argon with 2–5% hydrogen for flux-free oxide reduction before soldering.

Safety, compliance and validation paths in the US

For a US installation the binding constraints are ozone exposure limits, exhaust design, electrical safety listing and, for medical devices, FDA clearance — all of which are manageable if they enter the project at the specification stage rather than after installation.

Ozone, NOx and exhaust requirements

Any discharge in air generates ozone; discharges in air also generate nitrogen oxides, which is why an untreated exhaust can irritate airways and corrode ductwork. OSHA sets the permissible exposure limit for ozone at 0.1 ppm as an 8-hour time-weighted average, and nitrogen dioxide at 5 ppm as a ceiling. In practice that means enclosing the treatment zone, extracting at negative pressure and, for anything above modest duty cycles, adding catalytic or thermal abatement before discharge to atmosphere.

Under NFPA 79 and ANSI/NFPA 654 practice, the enclosure, interlocks and exhaust verification belong in the electrical and safety documentation, not in a vendor brochure. Ask for the ozone generation rate in grams per hour at your intended power and gas flow — a number every credible manufacturer can supply and many will not volunteer.

UL, CE, OSHA and FDA pathways

US installations typically require a UL 61010-1 or UL 508A listed panel and a machine that satisfies OSHA 1910 subpart O guarding requirements. CE marking under the EMC and Low Voltage Directives matters for exported equipment. For medical use, an atmospheric plasma device falls under FDA device regulation: plasma physics is well understood, but clinical claims require a 510(k) clearance or a PMA, with electrical safety reviewed under IEC 60601-1 and biocompatibility under ISO 10993.

"Plasma is often described as the fourth state of matter — a gas energetic enough that its atoms have lost or gained electrons." — Encyclopaedia Britannica, entry on the plasma state of matter.

Is cold atmospheric plasma safe for skin?

Within controlled parameters, yes. Cold atmospheric plasma is already used clinically for wound disinfection, chronic ulcer treatment and dermatologic resurfacing, and ClinicalTrials.gov lists 38 registered studies involving plasma devices. The safety envelope is defined by gas temperature, UV output, ozone in the room and treatment duration — which is why clinical devices carry strict dosimetry and why consumer claims deserve skepticism.

Throughput, line integration and defect troubleshooting

Corona treats faster than any jet-based atmospheric plasma system and slower than a well-designed DBD array, so the throughput decision usually resolves itself once you know your web width and required dose.

Speed, width and cycle time

Corona stations handle 30–400 m/min on webs up to 3 m wide with a single pass. DBD arrays for atmospheric plasma run 10–120 m/min on widths up to about 2 m, depending on dwell time and gas. APPJ heads are an order of magnitude slower per unit area — 0.2–5 m/min per jet — so they are rastered or arrayed, and robotic cells typically accept 5–30 seconds per part. For inline extrusion or printing at 250 m/min, corona or a wide DBD is the only realistic option.

Integration questions that decide the layout: does the surface need treatment immediately before the coating nip (yes, always — decay is real), how much space exists between the die and the first roller, and can the head be retracted for threading. A 300 mm gap between treatment and coating is comfortable; a 3 m gap is a quality risk.

A defect troubleshooting checklist

Most field complaints trace to four causes: over-treatment, contamination, gap drift or moisture. Work through this sequence before assuming the hardware is at fault.

1. Verify incoming substrate — measure dyne on untreated stock; a change in slip additive or lot can mask everything downstream.
2. Check gap and parallelism with a feeler gauge; a 0.5 mm drift on a DBD reduces dose by roughly 10%.
3. Test treatment decay by treating a sample, then re-measuring dyne at 1, 24 and 72 hours.
4. Inspect for pin-holes and bloom under 10× magnification — filament burns show as circular defects, thermal damage as gloss change.
5. Confirm gas purity and flow; a partially blocked nozzle changes plume chemistry before it changes plume appearance.
6. Log humidity. Above 60% RH, dyne readings drop 2–3 mN/m without any change in the process.
7. Reduce power in 10% steps if adhesion is fine but the surface looks hazy — over-treatment creates a weak boundary layer that fails cohesively.

How long does plasma activation last before it fades?

Untreated polymer surfaces are low-energy and stable; treated surfaces are high-energy and thermodynamically uncomfortable, so polar groups rotate inward and mobile additives migrate to the surface. Typical decay to half the original dyne gain runs 3–7 days for corona, 10–21 days for atmospheric plasma, and 14–30 days for low-pressure plasma. Sealed storage in a cool, dry area extends all three figures considerably.

How to choose: decision path, glossary and a real case

Choose corona when you have a flat, wide, non-conductive web, high line speed and a modest adhesion target; choose atmospheric plasma when you need low heat, gas chemistry control, conductive substrates or robotic 3D coverage; choose low-pressure plasma when penetration into complex geometry matters more than cycle time.

Glossary of terms you will hear from vendors

DBD (dielectric barrier discharge): an electrode pair with an insulating layer that prevents arc formation, producing a diffuse or filamentary discharge. APPJ: a nozzle-fed jet that projects plasma beyond the electrodes. CAP: cold atmospheric plasma, referring to near-room-temperature operation. Piezoelectric direct discharge: a transformer-driven source that generates plasma at the ceramic tip without a separate gas supply line. RONS: reactive oxygen and nitrogen species, the chemically active output responsible for both surface functionalization and antimicrobial action.

Case study: 12 µm PET before UV flexo printing

A Wisconsin converter running 12 µm PET at 60 m/min on a 1,400 mm web could not hold ink adhesion at the edges after switching suppliers. Untreated dyne measured 38 mN/m; corona raised it to 44 mN/m at the center but only 39 mN/m at the edges, and the film showed filament pin-holes on roughly 0.3% of the roll. A DBD atmospheric plasma array with compressed air and nitrogen was installed ahead of the print station at a capex of $145,000.

Results after commissioning: dyne stabilized at 56 mN/m across the full width with a standard deviation of 1.2 mN/m, water contact angle dropped from 68° to 24°, ink adhesion passed the 3M 610 tape test at 5B, and pin-hole defects fell below detection. Gas and power added $2.40 per operating hour; scrap reduction paid the installation back in 14 months. The process window that held was 1.5–2.5 mm gap, 55–70% power and under 55% RH — outside it, the edges drifted first.

US supplier landscape, lead times and leasing

The US market splits into corona specialists serving packaging converters, DBD integrators serving medical and electronics assembly, and APPJ vendors concentrated around semiconductor packaging and wire bonding. Standard lead times in 2026 run 6–12 weeks for corona and 10–20 weeks for atmospheric plasma systems with custom tooling; semiconductor-grade argon/hydrogen sources are longer. Options beyond purchase: leasing at roughly 2.5–3.5% of capex per month, demo units from integrators, and a thin secondary market in refurbished corona stations. Service networks matter more than price — a two-day response time beats a 5% discount on a line running three shifts.

Frequently asked questions

Conclusion

The difference between these two technologies is less about whether they work and more about what you are willing to trade. Corona gives you speed and low capital on flat webs; atmospheric plasma gives you low heat, controlled chemistry and access to substrates and geometries corona cannot reach. Neither is universally superior, and any vendor claiming otherwise is selling rather than engineering.

Start from the surface specification, not the hardware. Measure untreated dyne, define the adhesion test, then work backward to dose, gas and dwell time. When you do that, the atmospheric plasma decision usually makes itself — and the numbers in this guide give you a defensible basis for the capital request.