Washing Produce with Plasma-Activated Water: A 2021 Review of Reported Kills, Mild-Heat Pairings and Open Gaps (Soni, Choi and Brightwell)

This open-access review by Aswathi Soni, Jonghyun Choi and Gale Brightwell of AgResearch in New Zealand (received 18 December 2020, accepted 12 January 2021) gathers laboratory studies of Plasma-Activated Water (PAW) as a disinfectant, mainly as a wash for fruit and vegetables. Its overall verdict is favourable but hedged: the authors see "significant potential" for a non-chemical, non-thermal intervention, while noting that results depend on the food, the microbe and cost. Read for its numbers, the review describes a modest effect on produce, usually between about half a log and just over two logs, which is to say a reduction of roughly two-thirds to about 99 per cent of the counted microbes. That is far short of heat pasteurization.

What is plasma-activated water said to be?

The review defines it as water exposed to cold atmospheric plasma (see Non-Thermal Plasma (Cold Atmospheric Plasma)) under adjustable settings such as voltage, carrier gas, temperature, pulsing and frequency. It tabulates seven generator designs from the literature, ranging from bubbling electrodes and plasma jets to a rotating-nozzle jet and a DC-driven streamer corona with transient spark. The common outcome is acidification, higher conductivity and higher oxidation-reduction potential (ORP). Reported examples include pH falling from 6.8 to 2.3 with ORP rising from 250 to 540 mV, and an oxygen-argon jet run for 15 min that took pH from 7 to 3, ORP from 250 to 550 mV and conductivity from 0 to 410 μS/cm. The authors stress that no single setting explains these changes; activation time, power, temperature and bubbling all appear to contribute. A tap-water study elsewhere in this dossier shows that strong acidification is not universal (see What Plasma Does to Tap Water: Sixteen Measured Species, a Conductivity Check and a Lentil Trial (Judée et al., 2018)).

What kill figures does the review report on produce?

The produce table lists six studies:

The text summarises these as washing effects of up to 3 log, a figure the table itself does not reach. It also describes the strawberry result as below 1 log, whereas the table gives 2.0 log for the same reference. The authors concede that heat pasteurization achieves at least 6 log against non-spore-forming bacteria and that the two cannot be directly compared. Their stronger claim concerns shelf life. For refrigerated apple slices stored for 12 days, the review gives two figures for the same study: a 1.05 log CFU/g reduction in aerobic bacteria on day 12, and counts 2 log below the untreated sample at that point. It does not reconcile the two, so the effect is best read as roughly 1 to 2 log. On strawberries, no fungal hyphae appeared over six days of storage, and the review suggests the water reached crevices that protect microbes from ethanol and UV light. The review attributes the delayed regrowth to sub-lethal injury, an interpretation rather than a measured mechanism.

In simple liquids the reported numbers are larger. Several studies in buffers or broth reached more than 5 log, and one in saline buffer exceeded 7 log. Food surfaces, the authors note, change how microbes attach and resist.

Does pairing with mild heat help?

Most of the review's strongest figures come from combinations, which it calls a "hurdle" approach:

• Yeast in suspension: PAW plus 50 °C for 6 min gave 4.4 log, against 0.27 log for PAW alone at 25 °C and 1.92 log for heat alone.
• Yeast on grapes: PAW with heat at 50–55 °C for 30 min gave 2.4 log CFU/g.
• Salted kimchi cabbage: a 10 min PAW wash followed by 5 min at 60 °C gave 3.4 log for Listeria monocytogenes and 3.7 log for S. aureus.
• Bacillus cereus spores: combinations at 40 °C and 55 °C over 60 min gave 1.54 and 2.12 log CFU/g. Microscopy showed damage to the spore coat and cortex.
• Ultrasound: pairing with ultrasound gave 1.33 log for E. coli K12 and 0.83 log for S. aureus, against under 0.5 log for PAW alone. The difference is under 1 log. The device is described as running at 40 Hz, although the review elsewhere places ultrasound at 20 kHz to 1.5 MHz.

How is it supposed to kill?

The review describes oxidative stress. Reactive species, it says, damage membranes through lipid peroxidation, and this damage then extends to proteins and DNA once the cell's antioxidant enzymes are overwhelmed (see Reactive Oxygen and Nitrogen Species (RONS)). It credits short-lived radicals with immediate killing and longer-lived nitrate, hydrogen peroxide and ozone with suppressing the recovery of injured cells. The supporting example is a shrimp study, which the authors say did not establish the effect of any individual species. The review also passes on that study's suggestion that the choice of feed gas made no direct difference, and adds that this still needs testing before it can be treated as established.

What does the review say does not work, or is unknown?

• Biofilms: the evidence conflicts. One study reports more than 5 log against S. aureus biofilm. Others found PAW alone ineffective against E. coli K12 biofilms or weaker against attached cells than free-floating ones. The authors conclude it would need at least one other technology for biofilm cleaning.
• Lab-only evidence: the authors write that nearly all work is on produce surfaces in laboratories, and they identify a large gap in evidence from real supply chains such as packhouses and cool stores.
• Cost and residues: they call the energy and cost of effective treatment high. They also describe residual reactive species in stored food as underexplored.
• Sensory testing: no trained-panel sensory studies exist. Colour and firmness checks were mixed. Spinach and strawberries showed no significant change, while treated red grapes changed colour faster than controls.
• Spores: it is unknown whether PAW triggers spore germination. There are no D-value comparisons with heat and no Gram-positive versus Gram-negative comparison.

How it fits the dossier

This review compiles other researchers' results and adds no new experiment. Its tables mix units (CFU/mL and CFU/g) and often leave power or exposure time unreported. It also does not say whether the underlying studies chemically neutralized residual oxidants before counting. That is the step a 2026 paper argues can inflate PAW kill rates (see The Neutralization Problem: Why Plasma-Activated Water Kill Rates May Be Overstated). The review's food focus complements the dossier's agricultural thread at Plasma Agriculture. Its modest standalone figures can also be read alongside Adding Iodide to Plasma-Activated Water: Faster Kills, a Proposed Hypoiodous Acid Mechanism, and a Salmonella Exception (McClenaghan et al., 2025), a 2025 study that added iodide to the water to strengthen the effect against food pathogens.