Before a film is spin-coated, dip-coated or slot-die coated onto ITO or glass, most labs finish the clean with one of two dry steps: UV-ozone or oxygen plasma. Both remove the last nanometres of organic contamination, both leave a surface that water spreads on, and both are often described as interchangeable. They are not quite. They work at different pressures with different active species, they treat different faces of a part, they change ITO in related but not identical ways, and they differ in cycle time, hardware and safety. This guide compares the two step by step and ends with a decision table and a short side-by-side test on your own substrates.
In one paragraph: on lightly contaminated, precleaned ITO and glass both methods give a clean, high-energy surface, and for ITO both remove carbon and raise the work function [4][5]. Plasma can be faster, can reach accessible faces of complex parts and may support other gases, but it needs a vacuum system and adds ion bombardment, which at high dose can roughen a surface and, with argon, can strip oxygen from an oxide such as ITO [4]. UV-ozone needs no vacuum or gas cylinder, avoids ion bombardment and is simple to run in batches of flat substrates, but can be slower, works mainly on surfaces that face the lamp and needs ozone extraction. ITO surface properties can change during storage [6][9], so control the queue time and check stability under your own conditions.
1. The short answer
The UV ozone vs plasma cleaning choice comes down to a handful of properties. The table summarises them for the usual case: precleaned ITO-coated or plain glass about to receive a solution-processed film.
| Property | UV-ozone | Oxygen plasma |
|---|---|---|
| Pressure | Atmospheric, in air | Vacuum, typically a few Pa to about 100 Pa of process gas |
| Active species | UV photons, ozone, atomic oxygen made near the surface | Oxygen atoms and other radicals, ions, electrons, UV from the glow |
| Ion bombardment | None | Yes, at a level set by power, pressure and configuration |
| Typical time for a thin organic film | Minutes at benchtop gaps | Seconds to a few minutes |
| Which faces are treated | Mainly those facing the lamp | Accessible faces of complex shapes; uniformity depends on loading |
| Effect on ITO | Carbon removed, work function raised | Carbon removed, work function raised; surface oxygen species changed |
| Gas choice | Air (oxygen-containing atmosphere) | Oxygen, argon, nitrogen or mixtures, depending on the instrument |
| Hardware | Lamp, chamber, extraction | Vacuum chamber, pump, gas supply, generator |
| Main hazard | Ozone and short-wave UV | Vacuum, process gas handling, electrical power |
General comparison drawn from the cited reviews [1][3]; exact behaviour depends on the instrument, the process settings and the substrate. Times are orders of magnitude, not recipes.
For most ITO and glass work either gives a coatable surface; the details below decide which fits your process.
2. Two mechanisms
UV-ozone: chemistry driven by light
In a UV-ozone cleaner the 185 nm mercury line splits oxygen molecules, and the atoms join other O2 molecules to form ozone. The 254 nm line is absorbed by ozone, releasing atomic oxygen, and by many organic molecules, which it excites or breaks. Excited fragments react with the atomic oxygen and ozone and leave as CO2, H2O and other small volatile molecules. Light and ozone together are much faster than either alone [2]. The process runs at atmospheric pressure without plasma-ion bombardment; UV absorption, oxidation reactions and heating can all affect the surface. The UV-ozone guide in this series, UV-Ozone Cleaning Explained: 185 nm, 254 nm and What Reaches the Surface, covers this chemistry in detail.
Oxygen plasma: chemistry driven by electrons
In a plasma cleaner an alternating electric field accelerates the few free electrons in the low-pressure gas until they ionise and dissociate gas molecules in collisions. A steady glow forms in which electrons, positive ions, neutral oxygen atoms, excited molecules and ultraviolet photons all exist together [3]. The sample is in contact with this glow through a thin boundary layer, the sheath, where ions are accelerated towards the surface. Cleaning then has two parts: a chemical part, in which oxygen atoms and other radicals oxidise organic material to volatile products, and a physical part, in which ion impact breaks bonds and, for heavy ions, can knock atoms off the surface [3]. Glow UV also contributes, as it does in a UV-ozone cleaner.
| What acts on the surface | UV-ozone | Oxygen plasma |
|---|---|---|
| Ultraviolet photons | Yes, 254 nm and any 185 nm transmitted through the gas path | Yes, from the glow |
| Ozone | Yes, main reservoir of oxidant | Can form; abundance depends on discharge conditions |
| Atomic oxygen | Yes, made near the surface, mainly by ozone photolysis | Yes, main oxidant, made throughout the glow |
| Ions | No | Yes, accelerated across the sheath |
| Electrons | No | Yes |
Qualitative summary; the balance of species in a plasma depends strongly on gas, pressure, power and drive frequency [3].
Plasma adds ion bombardment to photochemical and chemical effects, providing another way to remove material or modify the substrate. How much depends on the instrument; drive frequency is one factor, covered in Benchtop Plasma Cleaners Compared: 25 kHz, 40 kHz and 13.56 MHz.
3. Pressure and what reaches the sample
The mean free path, the average distance a gas molecule travels between collisions, shows how differently gas behaves at the two pressures. For a simple hard-sphere gas it is λ = kT / (√2 π d2 p). Taking an effective molecular diameter d of 0.36 nm for O2 and room temperature:
| Pressure | Where it applies | Mean free path (approx.) | Compared with a 3 cm gap |
|---|---|---|---|
| 101 kPa (1 atm) | UV-ozone cleaner | 70 nm | About 400,000 times shorter |
| 100 Pa | Upper end of plasma cleaner pressures | 70 µm | About 400 times shorter |
| 10 Pa | Typical low-pressure plasma | 0.7 mm | About 40 times shorter |
| 1 Pa | Low end | 7 mm | About 4 times shorter |
Illustrative arithmetic with an assumed molecular diameter; real mean free paths differ between species, and charged particles behave differently. The 3 cm gap is close to the lamp-to-sample heights discussed in section 11.
Mean free path describes collisions, not reactive-species lifetime or treatment depth, but it sets how easily species spread: the diffusion coefficient scales with it, so at 10 Pa a radical diffuses roughly ten thousand times more readily than in air at atmospheric pressure. UV-ozone cleaning depends on light and oxidants reaching the surface. Plasma-generated neutral radicals can reach surfaces outside direct line of sight, but access and uniformity depend on geometry, loading and surface losses. Shielded contact areas and deep features require validation. The PT13M product page states that it can process complex shapes; as with any plasma cleaner, check treatment on the faces that matter.
4. Speed and depth
A widely cited comparison of gas-phase oxidation methods comes from work on polymer films rather than ITO. Strobel and co-workers treated well-characterised polypropylene and poly(ethylene terephthalate) films by corona, flame, remote air plasma, ozone and UV/ozone, and analysed them by XPS, contact angle and infrared spectroscopy [7]. They report:
| Finding | Corona, flame and remote air plasma | UV/ozone |
|---|---|---|
| Time to reach an XPS O/C ratio above 0.10 on polypropylene | Less than 0.5 s | Orders of magnitude longer exposure |
| Where the added oxygen sits | Concentrated near the outer surface; corona and plasma somewhat deeper than flame | Reaches farther into the bulk of the polymer |
Summary of the abstract of Strobel et al. [7]. These polymer-film results do not establish cleaning times or a depth ranking for low-pressure oxygen plasma versus UV-ozone on ITO or glass.
For ITO and glass, compare the time needed to reach the same measured cleaning end point. Include loading, pump-down, gas stabilisation, venting and ozone clearance when throughput matters.
In the studied polymers, UV/ozone modified more deeply than the compared treatments. Modification depth and damage depend on the polymer and treatment conditions; neither method is “gentle” on every material.
5. ITO: work function and how long it lasts
On ITO, surface treatment can change energy alignment with the next layer; the preferred work function depends on the device architecture. Four findings frame the comparison.
Both remove carbon and raise the work function
Sugiyama and co-workers measured ITO by ultraviolet and x-ray photoemission after different cleans [4]. After ultrasonic cleaning in organic solvent alone, considerable carbon remained and the work function was 4.5 eV. UV-ozone removed much of the carbon and raised it to 4.75 eV. They concluded that carbon removal is the main cause of the work-function increase in both UV-ozone and O2 plasma treatments. Milliron and co-workers found that oxygen plasma changed the ITO work function and electron affinity by about 0.5 eV, which correlated with a higher injected current in simple hole-only devices [5].
Plasma also changes the oxygen at the surface
Milliron and co-workers found that adding or removing surface hydroxyl groups could not explain the plasma effect. Photoemission showed a new type of oxygen species, and they proposed that surface Sn–OH is oxidised to Sn–O• units [5]. So in their study, plasma did more than remove carbon. Sugiyama’s argon-ion sputtering result shows the other side of the same point: Ar+ sputtering also removed carbon, but it removed oxygen from the ITO and lowered the work function to 4.3 eV [4]. A purely physical, inert-gas treatment is therefore not a substitute for an oxidising one on ITO.
The device effect can be large
Wu and co-workers showed what an oxygen-plasma-treated anode can do in a single-layer doped-polymer device: drive voltage fell from more than 20 V to less than 10 V, external quantum efficiency rose about fourfold, and device lifetime improved by two orders of magnitude [8]. Kim and co-workers compared chemical treatments and oxygen and argon plasmas on their ITO and found that oxygen plasma gave the highest efficiencies and luminances and the lowest drive voltages, with device efficiency tracking the ITO work function [9].
Stability depends on treatment and storage conditions
Yamashita and co-workers measured the work function of ITO in situ after UV/ozone treatment. It rose immediately after treatment, then fell with time in air and finally returned to its initial value [6]. The decay followed first-order kinetics, ran faster at higher temperature with an estimated activation energy of 22 kJ/mol, and was not affected by humidity. Kim and co-workers, by contrast, reported that the work-function changes from their oxygen-plasma treatment were unchanged after long-term storage in air in the dark [9]. The two studies used different ITO, treatments and measurement methods, so they do not prove that one method lasts longer than the other. They do show that the post-clean work function is not a fixed property and should be checked under your own conditions.
| Study | Treatment | Reported result |
|---|---|---|
| Sugiyama et al. 2000 | Solvent; UV-ozone; Ar+ sputtering | Work function 4.5 eV; 4.75 eV; 4.3 eV (oxygen lost) |
| Milliron et al. 2000 | Oxygen plasma | Work function and electron affinity changed by about 0.5 eV; new surface oxygen species |
| Wu et al. 1997 | Oxygen plasma (device study) | Lower drive voltage, about 4x efficiency, about 100x lifetime |
| Kim et al. 1998 | Chemical treatments; O2 and Ar plasma | O2 plasma gave the highest efficiencies on their ITO; work-function change stable in dark air storage |
| Yamashita et al. 2015 | UV/ozone, then air exposure | Work function returned to initial value over time; faster when warmer; humidity no effect |
Results of the cited authors under their own conditions, from the published abstracts [4][5][8][9][6]. ITO from different sources and different treatment settings will give different numbers.
In practice, either method can give a higher-work-function ITO surface with reduced carbon contamination; plasma adds surface-oxygen chemistry and, at high dose, a risk of changing the ITO itself. With either, coat after a short, fixed queue time and keep the ITO lot, preclean and dose constant within a study.
6. Glass, quartz and silicon
On glass and fused quartz the goal is simpler: remove the hydrophobic organic film so that a water-based or polar ink wets the surface evenly. Both methods do this well on precleaned glass. A common acceptance test is the water contact angle; ASTM C813 describes a contact-angle method for hydrophobic contamination on glass [10]. Effective treatment can produce a low water contact angle; the result and its stability depend on the glass and exposure conditions.
The differences show up at the edges of the process:
- Geometry. For a flat slide lying under the lamp, UV-ozone treats the upper face and leaves the lower face largely untreated. Plasma can treat both faces and edges if the holder leaves them accessible, which matters for dip coating.
- Patterned or textured surfaces. Plasma radicals can reach beyond direct line of sight, but treatment inside trenches and pores depends on geometry and surface losses. Validate coverage for either method.
- Silicon. Both methods can oxidise bare silicon and change its surface chemistry and wettability. If the native oxide must be controlled, treat the clean as part of the oxide history.
If an ink still beads up after either clean, the cause is often the ink or the time between clean and coat, not the clean. Why Your Ink Dewets: Surface Energy, Contact Angle and Wetting Before You Coat covers the wetting side.
7. Polymers and PDMS
On polymers both methods oxidise the polymer itself, not only the contamination on it.
| Aspect | UV-ozone | Oxygen plasma |
|---|---|---|
| Rate of surface oxidation | Depends on polymer, irradiance and oxidant delivery | Depends on polymer and plasma conditions |
| Depth of modification | Material- and dose-dependent | Material- and dose-dependent |
| Hydrophobic recovery | Occurs; depends on polymer and dose | Occurs; studied over more than 100 days for ten polymers, with clear differences between them |
| Typical use | Gradual, controlled oxidation; PDMS surface modification | Quick activation before bonding or coating |
From the cited abstracts [7][11][12]; behaviour depends strongly on the polymer and the dose.
Jokinen and co-workers treated ten polymers used in microfabrication, including PDMS, PMMA, polycarbonate, polystyrene and SU-8, with identical oxygen and nitrogen plasmas and followed their water contact angles for more than 100 days [11]. Recovery differed clearly between polymers and plasma settings. Efimenko and co-workers studied PDMS modified by UV and UV/ozone [12]. As for ITO, an activated polymer surface changes with time, so coat or bond on a fixed schedule.
8. Practical differences
| Factor | UV-ozone | Oxygen plasma |
|---|---|---|
| Cycle | Load, close, run, wait for ozone to clear, unload | Load, pump down, stabilise gas, run, vent, unload |
| Utilities | Power and ozone extraction | Power, vacuum pump, process gas, exhaust for the pump |
| Process variables | Gap, time, temperature, air exchange | Gas, flow, pressure, power, time, sample position |
| Three-dimensional parts | Faces away from the lamp clean slowly | Accessible faces treated; loading matters |
| Heating | Speeds cleaning; available on heated models | Sample warms with power and time |
| Main safety points | Ozone exposure, UV exposure | Gas cylinders, vacuum, pump exhaust, electrical power |
General comparison; follow the instrument manual and your institution’s safety rules for each item.
Heat works differently in each method. Vig reports that heating the sample speeds UV/ozone cleaning [1], so a hot plate is a useful lever. In a plasma the sample warms on its own, and for temperature-sensitive layers the risk is too much heating, not too little; shorter runs or lower power keep it down.
Safety also differs. Ozone is a respiratory irritant; NIOSH lists a recommended exposure limit of 0.1 ppm as a ceiling and an IDLH value of 5 ppm [13]. UV-ozone cleaners should run closed, with exhaust to extraction or an ozone destruction unit, and with the chamber cleared before opening. Plasma cleaners instead bring compressed gases, a vacuum pump and its exhaust, and a high-frequency power supply; oxygen service requires an oxygen-compatible pump and, if lubricated, an approved lubricant; check the manuals.
9. Which to choose
Where both fit, the side-by-side test in section 10 settles it.
| Situation | Usually fits | Why |
|---|---|---|
| ITO glass used as a hole contact, small batches | Either | Both remove carbon and raise the work function; control dose and queue time |
| Glass or quartz slides for spin or blade coating | Either; UV-ozone is simpler | Only the coated face needs treatment |
| Glass for dip coating | Plasma, or UV-ozone on both faces | Both faces and edges enter the solution |
| Three-dimensional or porous parts | Plasma | Radicals can reach beyond direct line of sight; validate deep features |
| Lab without vacuum line or gas supply | UV-ozone | Needs only power and extraction |
| High throughput, short cycle | Compare complete cycles | Treatment time alone excludes loading and utility overheads |
| Need for argon, nitrogen or mixed-gas treatment | Plasma | Gas can be changed on instruments rated for it; confirm at quotation |
| Polymer to be modified gradually | UV-ozone | Slower, easier to control by time |
| Quick polymer activation before bonding | Plasma | Often rapid; validate modification depth and dose |
| Oxide surface that must not be reduced or roughened | UV-ozone, or a mild oxygen plasma | UV-ozone has no ion bombardment and a mild oxygen plasma keeps it low; avoid argon sputtering |
Rules of thumb drawn from sections 2–8; they are starting points to test, not rules for every material.
Many labs keep both: UV-ozone for routine flat substrates, plasma for three-dimensional parts, polymer activation and other gases. Either is one step in a sequence of solvent cleaning, dry clean and coating. Substrate Cleaning Protocol for Thin-Film Devices: Solvents, UV-Ozone, Plasma and Queue Time covers the full sequence.
10. A side-by-side test
When the literature leaves the choice open, a short test on your own substrates and ink settles it:
- Split one batch. Take substrates from the same lot, preclean them together and divide them into three groups: no dry clean, UV-ozone and plasma.
- Run a short dose series for each method. For example three UV-ozone times and three plasma times, keeping the gap, gas and power fixed. The values are yours to choose; they should span from under-treated to clearly sufficient.
- Measure water contact angle within minutes of unloading, and again after a fixed delay such as one hour, to see how quickly each surface recovers.
- Coat all groups with the real ink after the same queue time and inspect coverage, edge pull-back and pinholes.
- Measure the device or film property that matters, such as adhesion, sheet uniformity or device efficiency, and pick the method and the lowest dose on the plateau.
Record the substrate lot, preclean, dose, queue time and room conditions with each result.
11. Levon™ UV ozone and plasma cleaners
The tables quote the Levon™ product pages; “Not stated” means the page does not give the value.
UV ozone cleaners
| Specification | UC100-SE | UC200-SE |
|---|---|---|
| Lamp | 300 W U-shaped lamp; 185 nm + 254 nm | 300 W lamp; 185 nm + 254 nm |
| Cleaning area | ≤8 in | ≤8 in |
| Lamp-to-sample height | 35 mm illumination height | 31–44 mm irradiation height (substrate thickness excluded), set by jacking |
| Heating | Not stated | 500 W hot plate; RT to 200 °C |
| Timer | Up to 99.99 h | Up to 99.99 h |
| Ozone concentration at the sample | Not stated | Not stated |
Plasma cleaners
| Specification | PT500 | PT13M | PT40K |
|---|---|---|---|
| Frequency | 25 kHz | 13.56 MHz, automatic matching | 40 kHz, automatic matching |
| Power | 0–500 W, adjustable | GE, SE: 0–300 W; BE: 0–200 W (maximum 300 W) | BE: 0–500 W (maximum 1,000 W); SE: 0–1,000 W |
| Chamber | Ø150 × 152 mm, aluminum alloy (larger sizes can be customized) | 226 × 126 × 224 mm, 6.4 L, aluminum alloy | 226 × 126 × 224 mm, 6.4 L, aluminum alloy |
| Vacuum level | ≤100 Pa | <10 Pa | <10 Pa |
| Gas control | Single gas path, 0.1–1 L/min | GE: dual MFC; BE: single float flowmeter; SE: dual float flowmeters | BE: single float flowmeter; SE: dual float flowmeters |
| Time setting | 1–200 s, continuously adjustable | GE, SE: 6,000 s; BE: 999 s | 300 s |
| Substrates | Not stated (chamber size above) | Under 8 inches; complex shapes | Up to 8 inches (200 mm) |
| Vacuum pump pumping speed | Not stated | GE, SE: 8 m3/h; BE: 4 m3/h | 4 m3/h |
| Gas stabilization time | Not stated | GE, SE: 30 s; BE: 40 s | 30 s |
| Process gases named | Not stated | Not stated | Not stated |
| Dimensions (W × D × H) | 452 × 386 × 225 mm; approx. 18 kg | 620 × 550 × 410 mm | 500 × 480 × 290 mm |
Specifications as published on the Levon™ UC100-SE, UC200-SE, PT500, PT13M and PT40K product pages on 22 October 2026. All five pages state that external utilities, where required, are supplied by the user. The supplied configuration is confirmed in the quotation.
How this connects to the guide: the UV ozone cleaners list illumination heights of 35 mm and 31–44 mm, close to the 3 cm gap used in section 3, and the UC200-SE adds the heat lever of section 8. The plasma pages list vacuum levels, not validated treatment-pressure ranges. Their flow-control configurations also do not establish compatible gases or mixture accuracy. Confirm operating pressure, gas compatibility including oxygen, and any mixing capability at quotation.
12. Keep Exploring the Surface Preparation Knowledge Hub
Start here
- Solution-Processed Thin Films: Choosing Spin, Dip, Slot-Die or Blade Coating — the coating methods that a clean surface prepares for.
- Plasma Cleaning: How It Removes Contamination — more on how plasma cleans.
- Plasma Surface Activation: Wettability and Adhesion — how treated surfaces gain wettability and adhesion.
- Plasma Ozone Generation: Inside the DBD Ozonizer — how ozone is made by discharge.
Cleaning and activation guides
- UV-Ozone Cleaning Explained: 185 nm, 254 nm and What Reaches the Surface — the UV-ozone chemistry in depth.
- Benchtop Plasma Cleaners Compared: 25 kHz, 40 kHz and 13.56 MHz — what the drive frequency changes.
- Substrate Cleaning Protocol for Thin-Film Devices: Solvents, UV-Ozone, Plasma and Queue Time — the full sequence from detergent to coater.
- Why Your Ink Dewets: Surface Energy, Contact Angle and Wetting Before You Coat — reading contact angles and fixing dewetting.
Links will be added as each article is published.
Levon™ UV ozone and plasma cleaners
- UC100-SE UV Ozone Cleaner — 300 W 185/254 nm lamp, 35 mm illumination height.
- UC200-SE Heated UV Ozone Cleaner — 31–44 mm adjustable height, hot plate to 200 °C.
- PT500 Compact Plasma Cleaner (25 kHz) — 0–500 W, Ø150 × 152 mm chamber, 1–200 s.
- PT13M RF Plasma Cleaner (13.56 MHz) — automatic matching, 6.4 L chamber, three gas-control options.
- PT40K Plasma Cleaner (40 kHz) — up to 1,000 W, 6.4 L chamber, substrates up to 8 inches.
- All Levon™ UV ozone cleaners and all Levon™ plasma cleaners — the category pages.
Questions about a specific substrate, gas or configuration can be sent through the contact form.
13. FAQ
Is UV-ozone or oxygen plasma better for cleaning ITO?
Neither is better in general. Both remove carbon contamination and raise the ITO work function; plasma can be faster and also changes the surface oxygen chemistry, while UV-ozone has no ion bombardment. Choose by geometry, throughput and available utilities, then fix the dose and queue time.
Is plasma cleaning faster than UV-ozone?
Plasma can be faster, but the cited sub-second results concern polymer oxidation by corona, flame and remote air plasma. They do not establish cleaning times for low-pressure oxygen plasma on ITO or glass. Compare measured end points and complete cycle times on your own substrates.
Can I use argon plasma instead of oxygen plasma on ITO?
Argon cleans mainly by ion impact. In one photoemission study, argon-ion sputtering removed carbon but also removed oxygen from the ITO and lowered its work function. For a high-work-function ITO surface, an oxidising treatment such as oxygen plasma or UV-ozone is the usual choice.
How long does the effect of UV-ozone or plasma last?
It depends on treatment and storage. UV/ozone-treated ITO lost its work-function increase in one study; oxygen-plasma-treated ITO retained it in another. Use a fixed queue time. Contact angle checks wettability, not work function; measure work function separately when it matters.
Does UV-ozone clean the back of the substrate?
Mainly the face towards the lamp is treated, because the atomic oxygen that does the work is made where the 254 nm light arrives. For both faces, turn the substrate and repeat, or use plasma with both faces accessible to the treatment.