The California Air Resources Board once tested seventeen household appliances that emit ozone. Air purifiers, an ozone laundry system, a refrigerator freshener. The strongest ozone source of all seventeen turned out to be a fruit and vegetable washer.
15.4 mg of ozone per hour. The refrigerator air purifier, for comparison, put out 1.6 mg/h, almost ten times less.
That is where this topic gets interesting. The question is not whether ozone kills bacteria. It does. The question is where the ozone goes after you close the lid.
First, honestly: ozone is permitted on food
This needs saying plainly, or the whole article turns into a witch hunt. There is a dedicated section of US regulation, 21 CFR 173.368, and the wording is unambiguous: ozone “may be safely used in the treatment, storage, and processing of foods, including meat and poultry.” Gaseous or aqueous phase, to Food Chemicals Codex specification, under good manufacturing practice. FDA approved it in 2001.
Industry does use it. In wash water sanitizers ozone sits alongside chlorine and peracetic acid, with a working range of 2 to 10 mg/L.
So the argument here is not about legality. It is about dose, about stability, and about your kitchen as a room with air in it.
What ozone does on a leaf
The NCCEH review collects the numbers for ozonated water. Lettuce, 4 ppm ozone, 68°F (20°C), two minutes:
| Organism group | Reduction, log CFU/g |
|---|---|
| Enterobacteriaceae | 1.3 |
| Psychrotrophs | 1.5 |
| Mesophiles | 1.7 |
Roughly what 100 ppm chlorine achieves under the same conditions. A log and a half, about 97 percent.
Now the pattern you have already met in the pieces on vinegar and hydrogen peroxide. In suspension, where the bacterium floats naked, 3 ppm of ozone delivers more than 5 log in thirty seconds against staph, E. coli, Pseudomonas and Enterococcus. On a leaf, over four times longer, you get 1.3 to 1.7.
A hundredfold difference, and it is not about ozone. It is about where the contamination sits: in stomata, in surface folds, under organic matter that quenches the oxidizer on the way in.
One more figure worth keeping. Spores of Aspergillus brasiliensis in that same suspension were not reduced at all. Not even after thirty minutes.
Head to head against tap water
Fishburn, Tang and Frank did in 2012 what this field usually lacks: they put consumer washing methods side by side and tested them on six produce items against three pathogens.
Ozone failed to beat plain tap water on any of the six.
And there is a second half to that result, more interesting than the first. The authors checked what remained alive in the wash water afterward. After dilute chlorine and after electrolyzed water they found nothing, not once, on any commodity. After ozone they recovered live Salmonella and Listeria in one replicate out of three for broccoli, cantaloupe and spinach.
So the water in the bowl stayed intermittently infectious. And transfer from one bad item to the rest of the batch is the main risk a wash actually carries, as covered in the piece on wash water.
Ozone does not stay in the water
A technical detail sellers do not advertise. Ozone is unstable and poorly soluble. Concentration in water halves, from 3.0 to 1.5 ppm, in thirty minutes.
A processing line handles this by generating ozone continuously and monitoring the concentration continuously. In a bowl on your counter the concentration is unknown to everyone, you included. The device does not tell you how much ozone is in the water right now, how much of it the soil on the carrots has already consumed, or how much is left by the end of the cycle. How production sites control residual ozone, degas the water and protect the operator is worked through in the wiki: ozonation, residual ozone and safety.
Now the room
Back to the measurement this article opened with. Zhang and Jenkins, Indoor Air, 2017, both at the Research Division of the California Air Resources Board.
Seventeen appliances, nine emitting measurable ozone. The sentence about the produce washer, verbatim:
Multiple cycles of use of one fruit and vegetable washer increased personal exposure concentrations by an average of 2550 ppb, over 28 times higher than the level of the 1-h California Ambient Air Quality Standard for ozone (0.09 ppm).
2550 ppb of personal exposure. Twenty-eight times the state’s one-hour air quality standard.
Set that against the limits various agencies have set:
| Agency | Limit |
|---|---|
| FDA, indoor medical devices | no more than 0.05 ppm (50 ppb) |
| OSHA, workplace | 0.10 ppm averaged over 8 hours |
| NIOSH | 0.10 ppm, never to be exceeded |
| EPA, outdoor air | 0.08 ppm averaged over 8 hours |
What ozone does in the lungs, EPA states without softening: “When inhaled, ozone can damage the lungs. Relatively low amounts can cause chest pain, coughing, shortness of breath and throat irritation.” The same page continues into worsening asthma and a reduced ability to fight respiratory infections.
One honest caveat. That EPA document is written about ozone generators sold as air cleaners, not about produce washers. Substituting one for the other would be dishonest. The connection is simpler than that: a produce washer releases ozone into kitchen air, and kitchen air is exactly what Zhang and Jenkins measured. The exposure limits apply either way, because the lungs are the same lungs.
One more line from that EPA page deserves quoting, because marketing copy routinely inverts it:
Some vendors suggest that these devices have been approved by the federal government for use in occupied spaces. To the contrary, NO agency of the federal government has approved these devices for occupied spaces.
How to read “kills 99.9% of bacteria”
99.9 percent is exactly 3 log. It sounds large until you ask three questions.
- Measured on what. In suspension ozone gives over 5 log in half a minute. On a leaf it gives 1.3 to 1.7 log in two. If the claim does not say the test ran on actual produce, it probably ran on water.
- Compared against what. The honest comparator is tap water for the same time with the same rubbing. Ozone did not survive that comparison in Fishburn.
- What is left in the water. Three log on the produce counts for little if live Salmonella is still in the bowl when you lower the next batch into it.
What to do instead
- Running water and mechanics. Rubbing with your hands, and a brush on firm produce, land in the same order of magnitude as a chemical soak, in seconds.
- Do not soak the batch in one bowl. Standing water without a sanitizer carries contamination from one bad leaf to all the others.
- Dry after washing. FDA recommends a clean cloth or paper towel, which removes a further share of surface bacteria.
- Leave bagged washed greens alone. The reasoning is in do triple washed claims mean anything.
- For higher-risk eaters, the answer is heat. No soak is a kill step. For pregnancy, older adults and anyone immunocompromised, cooking is the control.
Bottom line
- Ozone is permitted for treating food, 21 CFR 173.368, and industry uses it. The argument is about dose and about the room.
- On lettuce, ozonated water at 4 ppm for two minutes gives 1.3 to 1.7 log. The same ozone in suspension gives over 5 log in thirty seconds.
- Head to head on six commodities, ozone never beat tap water, and live pathogens turned up intermittently in the wash water afterward.
- Ozone in water halves in thirty minutes, and a consumer device does not show you the concentration.
- The California Air Resources Board measured a fruit and vegetable washer as the strongest ozone source among seventeen household appliances, 15.4 mg/h, with personal exposure averaging 2550 ppb.
- Fungal spores under these conditions are not touched at all.
The neighboring home methods are covered separately: baking soda, vinegar, hydrogen peroxide, ready-made sprays. What industry actually does with wash water is in wash water sanitizers. The wider picture sits in the leafy greens food safety hub. Ozone has also been tested as a substrate treatment for mushroom growing, where it failed to control green mold, and what that means for a home grower is in our guide to growing mushrooms at home.
