How Plasma-Activated Water Ages in Storage: Peroxide, Nitrite and Radicals over Three Months at Three Temperatures (Cingesar et al., 2025)

Most studies in this dossier test Plasma-Activated Water (PAW) soon after it is made. A paper by Ivan Karlo Cingesar and seven co-authors, with Tomislava Vukušić Pavičić as corresponding author (Molecules, accepted 24 November 2025), asks what happens next. The team made treated water under eight combinations of process gas, discharge geometry and voltage, then stored it at room temperature, 4 °C or −18 °C. They followed hydrogen peroxide and nitrite for 105 days and checked radical and antioxidant activity at day 0 and day 60. Their headline is that refrigerated samples held on to the long-lived species best. No microbes were tested, so the paper describes chemistry, not disinfecting power.

What was made, and how?

The paper reports that 500 mL batches of demineralised water were treated for 7 min in a high-voltage pulsed discharge system built with the Zagreb company Impel d.o.o. It ran at 120 Hz, with 2 µs pulses at 20 or 30 kV, between a titanium needle and a grounded electrode set 10 mm apart. In the "gas-phase" set-up the needle sat above the water; in the "liquid-phase" set-up both electrodes were submerged. Nitrogen or oxygen was fed in through a tube. These are variants of the discharges described on Non-Thermal Plasma (Cold Atmospheric Plasma). Peroxide was measured with a titanium reagent at 410 nm and nitrite with Griess reagent at 540 nm. Radicals and antioxidant activity were measured by electron paramagnetic resonance (EPR), a technique that detects unpaired electrons.

The paper reports significantly lower total and specific energy for oxygen runs than for nitrogen ones, and the authors' conclusion names oxygen gas-phase treatment as the least energy-demanding process. Their own energy table does not quite bear that ranking out (see below).

What happened in storage?

Hydrogen peroxide settled into a narrow band everywhere. By the authors' account, nitrogen gas-phase samples levelled out near 2 mg/L at room temperature and near 3 mg/L at 4 °C and −18 °C; oxygen samples settled around 2 to 2.5 mg/L. Their conclusion is that 4 °C gave the highest stable peroxide. The tables only weakly support that: at day 105 most temperature comparisons within a row share a significance letter, meaning no significant difference (for example 3.25 mg/L at 4 °C against 3.12 mg/L frozen for nitrogen gas-phase water at 20 kV).

Nitrite shows the clearer storage effect. Nitrogen gas-phase water at 20 kV kept at 4 °C fell from 1.16 mg/L at day 14 to 0.45 mg/L at day 105 (the tables begin at day 14, so they give no nitrite figure for freshly made water), while its room-temperature twin reached 0.00 and the frozen one was at zero from day 35. Oxygen liquid-phase plasma produced essentially no nitrite at all. The authors frame the gas choice as application-dependent: nitrogen if nitrite is wanted, oxygen if it should be minimal.

Redox and acidity did not follow the textbook pattern. Against 252.3 mV for untreated water, oxidation-reduction potential rose in gas-phase samples (up to 443.6 mV) but fell in liquid-phase ones (down to 194.5 mV). Oxygen-treated samples also showed higher, more variable pH. Table 11 lists the same potential for samples that differ only in storage temperature, so these values appear to describe treatment, not ageing. Surface tension fell in every sample; the lowest value, 67.13 mN/m, came from the nitrogen, 20 kV, liquid-phase sample, against about 73 mN/m for pure water. For surface tension, process gas was the only factor the authors found statistically significant (p = 0.0356); power and reactor phase were not.

Radicals and antioxidant activity

Freshly made water showed no antioxidant activity in an EPR assay with the radical DPPH, except one nitrogen gas-phase sample at 30 kV (20.90 ± 6.12% signal reduction). After 60 days at 4 °C, low activity had appeared in most samples, reaching 19.38 ± 4.00% for oxygen gas-phase water at 30 kV, while the initially active sample dropped to 4.85 ± 1.08%. The authors describe this as the first direct measurement of antioxidant activity in the treated water itself.

Using the spin trap DIPPMPO, which the paper says it applied to plasma-activated water for the first time, the team reports slightly stronger hydroxyl-radical signals at 60 days than in fresh water. It also reports that samples with weaker radical signals tended to show more antioxidant activity. The authors' explanation is that species with unpaired electrons are left over once peroxide and nitrite formation "saturate" after about 60 days.

What the paper does not establish

• No biological test. The paper suggests that samples stored at 4 °C, which settled at higher hydrogen peroxide levels, could remain "effective" for uses involving contact with surfaces or food. It does not show that stored water kills anything.
• The radical assay adds its own chemistry. The methods mix the sample with added hydrogen peroxide (0.35 mM) and iron chloride before trapping. The paper mentions "Fenton-type reactions" only as one possible route that speeds up peroxide breakdown under oxygen. It does not discuss whether the added peroxide and iron contribute to the hydroxyl signal it reports.
• The mechanism contradicts the tables. It says peroxide and nitrite rise for up to 60 days, but the nitrite tables mostly show steady decline.
• A thinly supported correlation. The antioxidant–peroxide link rests on a handful of samples without a reported statistical test.
• Confounded storage. Frozen samples were held in 50 mL tubes filled to 25 mL, the others in 150 mL bottles, so container and fill differ along with temperature.
• Slips in the text. Room temperature is given as 25 °C in one table and 20 °C in others, voltages are once written as "20 V or 30 V", near-zero nitrite values are attributed to the frozen samples in a passage about room-temperature storage, a correlation is reported as a negative "r²", and the conclusion calls oxygen gas-phase treatment the least energy-demanding even though Table 2 lists lower figures for the oxygen liquid-phase runs (total 6552 and 9828 kJ, specific 13,104 and 19,656 kJ/dm³ at 20 and 30 kV) than for oxygen gas-phase (7056 and 12,096 kJ; 14,112 and 24,192 kJ/dm³). The data are available only on request.

How it fits the dossier

The paper adds a time axis the kill-rate studies lack. The 2026 neutralization study (The Neutralization Problem: Why Plasma-Activated Water Kill Rates May Be Overstated) reports a 28% peroxide loss in 24 h at 4 °C in hard water. This paper's demineralised-water peroxide instead plateaus for months, a reminder that water composition and reactor design matter. The 2018 bacteriophage study (Singlet Oxygen as the Virus-Killing Agent: Plasma and Plasma-Activated Water Against Three Bacteriophages (Guo et al., 2018)) found activity largely kept after 10 days at 22 °C, but credited singlet oxygen, which this paper did not measure. The fall in redox potential in liquid-phase samples is a counter-example to the general picture of rising potential on Reactive Oxygen and Nitrogen Species (RONS).