Wash Water Sanitizers: Chlorine, PAA and Ozone Doses

A fresh-cut processing line does not wash salad in plain water. Chlorine, peracetic acid or ozone goes into the flume at a measured dose, with sensors on the tank and a log sheet on the wall. From that, a reasonable-sounding idea follows: if the plant disinfects greens with chemistry, the home cook should do a smaller version of the same thing in a bowl.

The idea fails at the first step. The sanitizer in that tank is not there to clean the leaf. It is there so the water does not carry contamination from one leaf to the whole batch.

That is not an interpretation. The industry says it in capital letters.

What a wash water sanitizer does and does not do

From the Commodity Specific Food Safety Guidelines for the Lettuce and Leafy Greens Supply Chain, 1st edition, April 2006, written by the industry itself from farm to table:

Wash water disinfectants can be very effective in eliminating free-floating or exposed microorganisms. However, chlorine and other wash water disinfectants are used in wash systems to prevent the potential for cross contamination, NOT to surface sanitize produce. The 2011 Jensen Farms case shows what that distinction costs: the melon washing line had a catch pan for a chlorine spray, the spray was never switched on, and the listeriosis outbreak that followed killed 33 people. The case is in our piece on Listeria in cantaloupe.

FDA puts the same point in a section heading of its guide for ready-to-eat fresh-cut produce, and the heading is literally “What an antimicrobial substance in wash water does and does not do.” The text underneath:

Water containing an antimicrobial substance can remove some pathogens from the surface of produce, but the antimicrobials do not significantly minimize pathogens that have attached to produce or that have been internalized in the tissue. Thus, a wash water process control is not a “kill step” for pathogens on produce.

Then comes the part that explains why the chemistry is in the tank at all. During washing, pathogens dislodged from contaminated produce survive in the water and spread to uncontaminated produce washed at the same time or later in the same water. Uncontrolled, one bad head of lettuce ruins a large lot. The antimicrobial exists to stop that transfer.

One dirty leaf and nine clean ones

The clearest demonstration ran in the Journal of Food Science in January 2025 (Gao et al., volume 90, issue 1, e17657). The design is almost cheeky in how directly it models a bag of salad.

Take one baby spinach leaf, inoculate it with E. coli to roughly 8 log per leaf, drop it into the wash with nine clean leaves, agitate for one minute. Then count three things separately: what survived on the dirty leaf, what is floating in the water, and what moved onto the nine clean leaves.

Read the top and bottom rows together. Sodium hypochlorite at 80 ppm left 5.1 log on the dirty leaf, which means it removed something like a tenth of what was there. Judged as a disinfection step, that is a failure.

The same chemical at the same dose cut transfer to the neighboring leaves from 4.8 log to 1.1, a factor of roughly five thousand. Judged by the job it was actually hired for, that is a complete success.

The dirty leaf stays dirty. The nine clean leaves stay clean. Three rows, and the entire purpose of wash water chemistry is on the page.

The dose gap that explains the whole industry

A review by Banach and colleagues (International Journal of Environmental Research and Public Health, 2015, volume 12, issue 8, pages 8658 to 8677) supplies the number that makes everything else make sense.

To inactivate microorganisms in the water, a free chlorine residual of 1 to 5 mg/L is enough. To decontaminate the produce itself, you need 20 to 200 mg/L or more. The gap runs from twenty-fold to two hundred-fold.

Water is easy because a bacterium in it is naked and exposed. A leaf is hard because the bacterium sits in stomata, in surface folds and inside biofilm, under a layer of organic matter that quenches the oxidizer on contact. The authors state the conclusion plainly: sanitizers should be used to maintain the quality of washing water and prevent cross-contamination, rather than as a last resort for produce decontamination.

Chlorine: 200 ppm, one minute, and some fussy chemistry

Chlorine is still the workhorse. The maximum permitted for food contact without a subsequent rinse is 200 ppm available chlorine, and a typical treatment is one to two minutes in a 50 to 200 ppm solution.

Ohio State University Extension put a number on what that buys on the leaf, in factsheet AEX-262: one minute of submersion at 75 to 200 ppm active chlorine commonly achieves a microbial reduction of about 1.3 to 1.7 log. An industrial line with correct chemistry, live sensors and trained operators removes roughly 95 to 98 percent. That is the same order of magnitude you get from running tap water and mechanical rubbing.

Chlorine is also temperamental. The active species is hypochlorous acid, and its share depends on pH: highest near pH 6, still adequate below pH 7.5. Sodium hypochlorite itself sits above pH 11, so processors dose an acidulant such as phosphoric acid into the tank. They cannot push the pH too low either, because chlorine gas starts coming off, so there is a floor as well as a ceiling. And organic load simply eats free chlorine, so the concentration sags through a shift unless it is topped up.

PAA: why processors keep switching to it

Peracetic acid, or PAA, is made by reacting acetic acid with hydrogen peroxide. Under 21 CFR 173.315 it is permitted in produce wash water up to 80 ppm.

Its advantage is indifference to organic matter. Chlorine dies in the juice from cut leaves; PAA keeps working. In the Gao study, at high organic load (COD of 2500 ppm against 300), 80 ppm PAA left 2.9 log on the inoculated leaf while hypochlorite left 5.3.

The industrial-scale comparison came from Petri and colleagues (Journal of Food Protection, 2021, volume 84, issue 9, pages 1592 to 1602), run on real equipment with iceberg lettuce, carrots and baby greens. For wash water hygiene and for the microbial quality of the finished product, PAA and chlorine performed the same. The difference showed up elsewhere.

Washing with PAA produced no chlorinated disinfection by-products in any scenario. Washing with chlorine at 80 mg/L produced significant amounts of trihalomethanes, chlorates and chlorites. The chlorates and chlorites stayed under drinking water limits. The trihalomethanes exceeded the legal limit. Perchlorates were below the quantification limit throughout.

Ozone: flawless on paper

FDA approved ozone as an antimicrobial agent for the treatment, storage and processing of foods back in 2001 (21 CFR 173.368). On paper it is the ideal candidate: a powerful oxidizer that leaves no residue, decomposes to oxygen, is generated on site, and spares the plant from trucking and storing hazardous chemicals.

In suspension it lives up to the billing. Tested against European standards EN 1040 and EN 1275, ozonated water at 3 ppm inactivated Staphylococcus aureus, E. coli, Pseudomonas aeruginosa and Enterococcus hirae by more than 5 log in thirty seconds.

On lettuce the numbers change. A two-minute dip in ozonated water at 4 ppm and 68°F (20°C) gave 1.3 log against Enterobacteriaceae, 1.5 against psychrotrophs and 1.7 against mesophiles. Comparable to 100 ppm chlorine under the same conditions. One to two log again, exactly as in every section above.

Ozone’s real problem is practical. It is unstable and poorly soluble: concentration halves from 3.0 to 1.5 ppm in thirty minutes. A tank that was dosed and left alone stops being protected well before the shift ends. How ozone behaves once it is in a nutrient solution, and why growers pair it with UV-C, is covered in the wiki on ozone, UV-C and hydroxyl radicals.

The consequence is visible in Fishburn, Tang and Frank (Food Protection Trends, 2012, volume 32, issue 8), who checked the residual wash water after every treatment. After dilute chlorine and after electrolyzed oxidizing water, no viable pathogens were recovered at all, across six produce items. After ozone, viable Salmonella and Listeria turned up in one of three replicates for broccoli, cantaloupe and spinach. The single job the sanitizer is there to do is the job ozone does not always finish.

The three side by side

Look at the second row. Three different chemistries, three different mechanisms, three different price tags, and all three land on the same order of magnitude on the leaf. That is not coincidence. It is the ceiling built into the idea of washing.

Why none of this transfers to a kitchen

• There is no batch at home. A sanitizer saves a lot, not a portion. When a metric ton of greens moves through a flume every hour, there is something to save. When you rinse one bunch for one dinner, there is nothing for the contamination to spread to.
• 200 ppm is a ceiling, not a recipe. It is the upper limit above which a rinse becomes mandatory. It was written for a line with a sensor and a log sheet, not for a bowl where nobody measures anything and where the concentration collapses on contact with the first head of lettuce.
• Household bleach in a kitchen is a hazard. It is not a food product, the dose is guesswork, and the vinegar bottle is usually within arm’s reach. Acid plus chlorine releases chlorine gas.
• The target is out of reach anyway. Even a perfectly executed industrial treatment takes off 1 to 2 log. If an infectious dose of a pathogen is on those greens, 1 to 2 log will not save you. Which is precisely why the industry refuses to call washing a kill step.

One thing does carry over, and it is the same principle read backwards. If shared water spreads contamination on a processing line without a sanitizer, it does exactly the same thing in a bowl on your counter, where there is no sanitizer at all. So rinse under running water instead of soaking, and keep the whole batch out of one basin.

Bottom line

• A wash water sanitizer treats the water, not the produce. The industry prints it in capitals; FDA puts it in a section heading.
• 1 to 5 mg/L of free chlorine kills bacteria in the water. Decontaminating the leaf itself needs 20 to 200 mg/L or more.
• Chlorine, PAA and ozone all land on 1.3 to 1.7 log on the leaf. Washing is not a kill step in a plant, and it is not one at home.
• PAA beats chlorine on tolerance to organic load and on by-products, not on killing power.
• Ozone halves in concentration within thirty minutes, and pathogens occasionally survived in residual ozone wash water. Chlorine and electrolyzed water never showed that.
• Reproducing any of it in a kitchen is pointless: nothing to protect, no way to measure the dose, and the same low ceiling on the result.

This is the reason consumer produce-washing products keep returning the same modest numbers regardless of what is in the bottle. The individual cases are covered separately: vinegar, and where the two contradictory numbers about it came from, baking soda, which turns out to be about pesticides rather than microbes, and veggie wash sprays, which failed to beat running tap water in a single combination tested. The wider risk picture sits in the leafy greens food safety hub. The same question asked about appliances is covered in the pieces on ozone produce washers, UV-C wands and ultrasonic cleaners. Whether growing the leaves yourself sidesteps the question is covered in the piece on countertop hydroponic gardens. Why some cells are beyond the reach of water at all is covered in the piece on internalization.