UV-C Dose and D90 Values: How Much UV-C Kills Microorganisms?
· 6 min read
Every UV-C design comes down to one question: does enough dose reach the target? This guide explains how dose is measured, what a D90 value is, how to turn a desired reduction into a dose, and why the organism you design for matters more than almost anything else.
Irradiance, time and dose
Irradiance is the UV-C power arriving on a surface, usually given in µW/cm² or mW/cm². Dose, also called fluence, is irradiance multiplied by exposure time:
dose (mJ/cm²) = irradiance (mW/cm²) × time (s)
The same dose can come from a strong lamp for a short time or a weak lamp for a long time. That is why surfaces that are irradiated continuously, such as HVAC cooling coils, can be protected with modest irradiance, while air passing through a duct needs high irradiance to receive a useful dose in a fraction of a second.
Log reductions
Disinfection results are expressed as log reductions: each log is a further factor of ten.
| Log reduction | Share inactivated | Survivors from 1,000,000 |
|---|---|---|
| 1-log | 90 % | 100,000 |
| 2-log | 99 % | 10,000 |
| 3-log | 99.9 % | 1,000 |
| 4-log | 99.99 % | 100 |
What a D90 value is
The D90 of an organism is the UV-C dose that inactivates 90 % of it, a 1-log reduction. With simple first-order kinetics, each further log needs the same dose again, so:
dose for n-log reduction = n × D90
For example, an organism with a D90 of 3 mJ/cm² needs about 9 mJ/cm² for a 3-log (99.9 %) reduction. Designers often add a safety factor on top, for example doubling the dose, to cover uncertainty in the data and in the installation.
Typical D90 values at 254 nm
The values below are representative mid-range figures from published compilations, including Kowalski’s Ultraviolet Germicidal Irradiation Handbook (2009) and the IUVA review by Malayeri et al. (2016). Reported values vary several-fold between studies, so treat them as design inputs, not guarantees.
| Organism | Type | D90 (mJ/cm²) | 3-log dose (mJ/cm²) |
|---|---|---|---|
| SARS-CoV-2 | Virus | 1.5 | 4.5 |
| Influenza A | Virus | 2 | 6 |
| Escherichia coli | Bacterium | 3 | 9 |
| Legionella pneumophila | Bacterium | 3 | 9 |
| Salmonella enterica | Bacterium | 3 | 9 |
| Listeria monocytogenes | Bacterium | 3.5 | 10.5 |
| Staphylococcus aureus | Bacterium | 3.5 | 10.5 |
| Pseudomonas aeruginosa | Bacterium | 4 | 12 |
| Mycobacterium tuberculosis | Bacterium | 5 | 15 |
| Bacillus subtilis spores | Bacterial spore | 20 | 60 |
| Penicillium spores | Mould | 40 | 120 |
| Aspergillus niger spores | Mould | 100 | 300 |
The spread is the key point: mould spores can need 30 times the dose of common bacteria. Designing for Listeria when the real problem is mould on a cold-store wall would under-size the system badly.
From dose to disinfection time
Divide the target dose by the irradiance at the point you care about. Using the weakest-lit point gives the time after which the whole surface has reached the target.
- Listeria, 3-log: 10.5 mJ/cm² at 200 µW/cm² (0.2 mW/cm²) takes 10.5 ÷ 0.2 = 52.5 seconds.
- Aspergillus niger, 3-log: 300 mJ/cm² at the same 200 µW/cm² takes 1,500 seconds, or 25 minutes.
Limits of the simple model
- Shoulders and tails. Some organisms show little effect at low doses before the curve steepens, and a small resistant fraction can survive longer than first-order kinetics predict.
- The medium matters. Values measured in water, in air and on surfaces differ, and humidity changes airborne results.
- Shading. UV-C travels in straight lines. Organisms in crevices, under dirt or inside biofilm receive much less dose than an open surface.
- Strains differ. Two strains of the same species can respond differently.
Pick the most UV-resistant organism you actually need to control, use the weakest-lit point rather than the average, and apply a safety factor. Then confirm the result on site, for example with swab tests before and after.