Quick answer
The physics of UV dose is the same in every medium — fluence is still irradiance × time, and every organism still has its dose-per-log. What changes completely, from air to surface to water, is how the dose gets delivered to the pathogen, and therefore what limits it.
- Water is the mature case: the pathogen flows past lamps in a designed reactor; the master variable is how transparent the water is.
- Air delivers dose on the move: microbes are irradiated during the brief time they spend in an UV zone; the master variable is air mixing and residence time.
- Surface is line-of-sight: only what the lamp can "see" gets a dose; shadows and distance dominate, and UV-C barely reflects to help.
Treating these three as "the same UV-C product" is how installations fail. Below, the delivery problem and the design lever for each.
1. The shared enemy: dose distribution
Across all three media the same thing decides success: the minimum dose any pathogen actually receives, not the average and certainly not the lamp's nameplate. A real system has a dose distribution across its volume or surface (validation exists precisely because of this), and the weakest spot governs the result. Each medium simply loses dose in a different way.
2. Water — the mature case
In water, the pathogen is carried through a reactor past the lamps, so the design problem is well-bounded — which is why water UV has reactors, RED validation and standards while air and surface largely don't (Section 5).
- The master variable is UV transmittance (UVT). Low-UVT water absorbs UV before it reaches the microbes; the average intensity a reactor delivers is computed from reactor dimensions, lamp number/output, lamp spacing, quartz-sleeve diameter, and water transmittance. Cloudy or coloured water can starve a perfectly good lamp.
- Lamp choice — LP vs MP. Low-pressure (LP) mercury lamps emit essentially one monochromatic line at 253.7 nm, very efficiently at the germicidal peak. Medium-pressure (MP) lamps emit a broad polychromatic spectrum across ~220–300 nm: much higher power density (fewer lamps, compact reactors) but lower electrical efficiency.
- Fouling fights back. Mineral scaling on the quartz sleeves cuts transmittance over time and worsens as sleeve temperature rises; serious reactors use wiper or air-scour cleaning, and chemical-plus-mechanical cleaning restores specific UV transmittance to >90 %, better than mechanical alone.
3. Air — disinfection on the move
In air, microbes are not held in front of a lamp; they are irradiated during the short time they pass through an UV zone. So the useful metric isn't a single dose but the equivalent air changes per hour (Eq ACH) the system adds — pathogen inactivation counted as if it were extra ventilation. A useful anchor: when UV inactivates 63 % of airborne organisms, that equals one equivalent air change; well-designed upper-room systems have been credited with up to ~24 Eq ACH.
Two dominant architectures:
- Upper-room GUV. A controlled germicidal zone is set up above head height; room air must be mixed vertically (occupant convection, ceiling fans, vents) so that lower air keeps cycling through the irradiated band. Without mixing, the upper zone disinfects air nobody is breathing.
- In-duct / AHU. Lamps inside ductwork or air-handling units irradiate the stream "on the fly". Here residence time is the lever — longer duct runs and lower air velocity raise the delivered dose, traded against pressure drop and fan energy.
The common failure is ignoring airflow and mixing: the lamp can be perfect and the room still under-treated because air doesn't actually reach the UV zone.
4. Surface — line-of-sight is everything
Surface disinfection is the least forgiving, for two physical reasons that no lamp upgrade fixes:
- Inverse-square falloff. Irradiance drops with the square of distance — double the distance, and only about a quarter of the irradiance remains. A floor far from a tower lamp gets a tiny fraction of what a nearby surface gets.
- UV-C barely reflects. Most building materials reflect UV-C poorly, so you cannot rely on bounce to reach shadowed faces. Anything not in direct line-of-sight — undersides, gaps, the back of a handle — receives almost nothing.
Consequently the delivered surface dose depends on room layout, source position, distance, surface angle, shadowing, lamp degradation and even humidity, which makes it genuinely hard to predict. The practical levers are line-of-sight to every target surface, distance, and longer cycle time for shadowed areas (or repositioning / multiple emitter positions). Measuring delivered dose at several points — not trusting a single timer — is what separates real surface disinfection from theatre.
5. Why water is "solved" and air/surface aren't
This asymmetry is worth stating plainly because vendor marketing blurs it. Water UV has decades of reactor engineering, RED biodosimetry and formal validation standards (see validation — UVDGM, DVGW W294, ÖNORM M 5873). Air and surface UV-C are effective when designed well, but their dose fields are far harder to control and the validation landscape is much weaker — there is no equivalent of a third-party RED certificate for a room-disinfection robot or an upper-room fixture. That doesn't make them invalid; it makes independent measurement and conservative design even more important there, because you can't lean on a certificate.
6. Choosing and specifying per medium
| If your problem is… | The delivery lever to specify | The trap to avoid |
|---|---|---|
| Waterborne (process/drinking/RAS) | UVT-rated reactor, LP vs MP for the duty, sleeve cleaning | low-UVT water starving the lamps |
| Airborne (rooms, HVAC) | Eq ACH target, air mixing, residence time | irradiating air that never circulates |
| Surface (rooms, equipment, conveyors) | line-of-sight coverage, distance, cycle time per zone | shadows + inverse-square + assuming reflection helps |
In every case, close the loop with the previous two articles: design the dose for the most resistant organism, and measure/validate the delivered dose rather than trusting the nameplate.
Cross-references
- UV-C Dose & Log Reduction — /knowledge/uv-c-dose-and-log-reduction: the dose physics shared by all three media.
- UV-C Validation & Dose Measurement — /knowledge/uv-c-validation-and-dose-measurement: why water has certificates and air/surface need measurement.
- Together these three form the UV-C foundation set. Forward: per-medium deep-dives (in-duct sizing, upper-room layout, surface-robot coverage) can branch from here.
Sources
- ACHR News — Upper-Room UV-C: A Proven Engineering Solution for Indoor Air Quality — equivalent air changes, upper-room design.
- AMCA — UV-C for HVAC Air and Surface Disinfection — in-duct dose = intensity × time × susceptibility.
- US DOE (Nardell) — GUV Energy Implications vs. Increased Ventilation — Eq ACH framing, 63% ≈ 1 air change.
- UV-C decontamination of a hospital room: amount of UV light needed (ScienceDirect) — surface dose distribution, geometry.
- Germicidal UV-C for Surface Disinfection in Hospitals — mapping the evidence (MDPI) — devices, parameters, effectiveness.
- Suez Water Handbook (Degremont) — UV disinfection sizing principles — reactor intensity from UVT, lamp output, sleeve, spacing.
- BC Gov — Guidelines for Ultraviolet Disinfection of Drinking Water — LP vs MP, water-reactor regulatory overview.
- Model of Radiation Transmittance by Inorganic Fouling on UV Reactor Lamp Sleeves (ResearchGate) — sleeve fouling, SUVT recovery.