1 INTRODUCTION

Non-thermal plasmas (NTPs), defined as ionized gases in non-equilibrium conditions, offer several technologies, processes and products for medicine and biology, among several other fields.[1] Besides many applications in surface modification processes of biomaterials,[2-7] recently NTPs emerged as newer tools in life sciences for clinical therapies in oncology and wound healing,[8] due to their ability to intervene in the redox control of oxidative diseases like cancer[9, 10] or infections.[9, 11] Therapeutic effects of NTPs arise from their ability to generate gas admixtures of reactive oxygen and nitrogen species (RONS) in air, close to living matter.[12, 13] RONS, well known as natural redox modulators, are involved in the stimulation or inactivation of many cellular functions such as growth, adhesion or death.[14]

Plasma medicine, hence, comes up with the idea to exploit NTP processes to exogenously deliver RONS to living matter to control cellular functions for therapeutic purposes.[9, 15, 16] It is important to mention that direct NTP treatments are always mediated by the presence of liquids like blood or exudate, water solutions whose components (i.e., inorganic salts, organic molecules and cells) can actively contribute to the further benefits or detrimental effects observed. These effects can be also obtained by using Plasma-Treated Water Solutions (PTWS),[17] namely, water-based liquid media enriched with RONS mixtures by NTPs ignited in N2/O2 mixtures with or without the presence of an inert gas like He or Ar. The Biological effects of PTWS are similar to those of direct NTP treatments, and could be used in clinics for wound healing,[18] cancer treatments[19-22] and disinfection.[23-26] Due to the possibility to be administered to various districts of the body, including those not easily accessible to plasma sources, PTWS-based therapies may become even preferred to direct NTP treatments.

The original composition of the liquid and the presence of water vapour have to be considered, to better elucidate the nature of such systems. Reactions with various molecules in the liquid generally extinguish transient species (primary RONS) like •OH, O2−, •O, N2* and others to form stable (secondary) RONS like H2O2, NO2− and NO3− in the PTWS.[27-29] These reactions, moreover, dynamically change the chemical composition of the treated liquids in a complex way, which strongly depends on their initial composition.[30, 31] In most in vitro experiments, then, the liquid used for RONS delivery must match the conditions for sustaining cell cultures, generally demanding a buffered pH and high concentrations of electrolytes and nutrients (amino acids, vitamins, growth factors, glucose). For these reasons, in simple cases, PTWS are generated from physiological or Phosphate-Buffered Saline (PBS) solutions;[32-34] much more frequently, indeed, cell culture media are preferred,[21, 35-38] with Dulbecco's Modified Eagle Medium (DMEM) and Roswell Park Memorial Institute (RPMI) medium among the most commonly used.[39]

The use of cell culture media allow studying at the best the effect of exogenous RONS delivery to cells, as their composition is optimized for culturing cells in vitro. Moreover, their chemical composition is similar to that of biological fluids like blood or exudates present during direct NTP exposure of tissues, whose alteration has never been studied in depth. Typical cell culture media generally, count more than 45 different components in high concentrations (from mg L−1 to g L−1) among inorganic (chlorides, metal ions, carbonate or phosphate buffers) and organic molecules (amino acids, vitamins, carbohydrates) plus optional additives like antibiotics and specific growth factors. The presence of these molecules makes the chemical composition of PT-culture media quite hard to be defined after NTP treatments, even though their biological efficacy is clearly evident. The study is also complicated by the need of using different media depending on the specific cell line under investigation.[20, 40] Quite frequently, for example, the effect of PTWS exposure is evaluated on many cell types in the same experiment, and the medium used to generate PTWS is changed depending on the cellular model under investigation. As it is often assumed that RONS enrichment in the liquids is exactly the same as long as the same plasma treatment is performed, in several cases the medium used for PTWS is not even specified, as highlighted by E. Biscop et al.[40] This assumption, indeed, is wrong in most cases.

Relevant compositional differences among different cell culture media are found in the concentration of buffer or antioxidant species,[40, 41] which are very likely involved in the scavenging of plasma-generated RONS. Therefore, overlooking the composition of the liquid may undermine the validity of in vitro results, as different cell types may be actually exposed to PTWS with a very different chemical composition resulting from the same NTP treatment. This aspect, combined with the scarce homogeneity of the NTP sources generally confines the validity of the biological effectiveness of PTWS within the specific lab in which they are produced, or leads to misleading comparisons of cell responses. This could be easily avoided by accurately evaluating the chemical composition of each PTWS, especially when they are generated from different culture media.

Predicting the chemical composition of PT-culture media, though, is hard to achieve. The complexity of the specific plasma–liquid interactions, in fact, must be added to the complex panorama of chemical reactions in the plasma phase. Most of the modelling studies concerning the mechanism of RONS formation in PTWS have been performed in liquids of simple composition like water,[25, 26, 28, 42, 43] by assuming plasma-driven processes supposed valid also in other liquids. In contrast, recent evidence attest that the formation of RONS in PTWS is heavily affected by the initial composition of the liquid. Regarding •NO, for example, we have recently reported that, after the same plasma treatment, •NO was revealed in the water while it was not in DMEM, partly for its scavenging by organic compounds in the medium and partly for the inhibition of formation routes at the buffered neutral pH of the medium.[44] In the same direction, Yan et al.[45] have found that H2O2 plasma-produced in DMEM was consumed, during storage, by reactions with specific amino acids such as cysteine and methionine, while the consumption was prevented by adding 3-nitro-tyrosine in DMEM, without specifying the ‘protection’ mechanism exerted by these molecules toward H2O2.[45] Privat-Maldonado et al.,[46] instead, reported the overproduction of long-lived RONS like H2O2 and NO2− in PT-cell culture media with respect to PT-water and PT-PBS after the same plasma treatment.[46] It is quite clear, thus, that biomolecules may affect the final RONS distribution in PTWS not only through scavenging processes but also through the promotion/inhibition of specific RONS production pathways in the liquid phase. In this paper, this last aspect is extensively discussed.

For this purpose, we have monitored the amounts of H2O2 and NO2− in PTWS generated from four liquids with very different chemical compositions: water, PBS and two cell culture media, DMEM and RPMI. Each liquid was exposed to Dielectric Barrier Discharges (DBDs) ignited in a wide range of operating conditions, by changing feed gas composition, treatment time and plasma energy dose. The contribution of each phase, the liquid and the plasma, to RONS generation, is analyzed and discussed. The data reported demonstrate that specific biomolecules in cell culture media effectively create alternative routes of RONS formation, which cannot be related to the action of plasma alone, above all in the case of H2O2. We, therefore, consider the results reported here of extreme importance for understanding the dynamic evolution of the chemical composition of PTWS. We propose to optimize the composition of each single PTWS for each specific biological protocol to be used in experiments and in therapies.

2 EXPERIMENTAL SECTION 2.1 PTWS generation

PTWS were generated with a DBD PetriPlas+ plasma source engineered by the Leibniz Institute for Plasma Science and Technologies (INP), and properly modified by E. Sardella in collaboration with INP researchers. A schematic overview of the apparatus is reported in Figure 1. The source consists of a Plexiglas flow unit, set with gas connections, and a discharge unit. This latter is composed of a stainless-steel ground grid 4 mm far from a disc-shaped high-voltage (HV) copper electrode (3 cm dia), covered with a glass dielectric disc (1 mm thick). A dry diaphragm pump (Pfeiffer Vacuum; Aßlar) is connected to the flow unit to evacuate the exhaust and keep the pressure constant at 760 Torr, as measured with an MKS Baratron. An electric field (6-kHz frequency) at different applied peak-to-peak voltages (Vpp) was generated with a power supply connected to a programmable 10 MHz DDS function generator (TG1010A; Aim-TTi) to ignite a volume discharge as wide as the HV electrode. The discharges were pulsed with a 25% duty cycle (DC; 25 ms plasma on, ton) over a period (t = ton + toff) of 100 ms.

Experimental setup to treat the liquids: Schematic overview of the plasma source. HV, high-voltage; MFC, mass flow controller; RONS, reactive oxygen and nitrogen species

The discharge unit is properly adapted for the remote treatment of liquids in commercial TPP® Petri dishes (57 mm dia, Techno Plastic Products; Trasadingen) housed under the reactor by means of a slot of the same diameter on the flow unit. When the dish is positioned, the gap between the liquid and the discharge becomes a closed system; purging the gap with the gas feed before igniting the plasma allows controlling the chemical composition in the gap. The treatment is achieved by the diffusion of RONS from the plasma to the liquid through the grid. Pure O2 (air liquid, 99.999%), N2 (air liquid, 99.999%), synthetic air (air liquid, 99.999%) and different O2/N2 mixtures were used as feed.

The gas flow rates were controlled with MKS mass flow controllers and kept constant at 0.5 L min−1.

The influence of the operating conditions thought to impact RONS concentrations in PTWS was extensively analyzed; treatment time, gas feed, applied voltage, treated liquid volume and treatment distance were varied to investigate the influence of each parameter on RONS concentration in the resulting PTWS. The ranges of each parameter are reported in Table 1. Detailed experimental conditions are reported in Section 3.

Table 1. Parameters changed to tune the amount of H2O2 and NO2− in PTWS Parameter Variation Liquid DMEM, RPMI, PBS, Water Gas feed O2, N2, O2/N2, synthetic Air Treatment time 60−300 s Applied voltage (Vpp) 10−14 kV Treated volume 2−5 ml Treatment distance 2−5 mm Abbreviations: DMEM, Dulbecco's modified Eagle medium; PBS, phosphate-buffered saline; PTWS, plasma-treated water solution; RPMI, Roswell Park Memorial Institute.

The discharge unit is housed in the Plexiglas flow unit with metallic positioner screws (Figure 1) designed to change the distance of the discharge unit from the surface of the liquid, regardless of the volume of liquid. This approach was used in the series of discharges at variable liquid volume (2–5 ml), and at a variable distance (2–5 mm) with a fixed liquid volume (2 ml).

The generation of RONS was analyzed in four different liquids: double-distilled water, PBS (cat. N: P3813; Sigma-Aldrich), DMEM (cat. N. D1145; Sigma-Aldrich) and RPMI (cat. N. R7509; Sigma-Aldrich). In some of the experiments, plasma treatments were performed also in l-cysteine (cat. N. C8755; Sigma-Aldrich) or l-tyrosine (cat. N. T-3754; Sigma-Aldrich) solutions, and in a minimum essential medium (MEM) amino acids solution (cat. N. 11140050; Gibco®, Thermo Fischer Scientific), containing the same non-essential amino acids of the standard MEM. The MEM amino acids solution was diluted 1:100 with water. Cell culture media were used without phenol red, to exclude spectroscopic UV-Vis interferences during the colorimetric detection of RONS, and without any of the supplements (i.e., fetal bovine serum or antibiotics) used in cell culture routines. The pH of the liquids was measured with a pH meter (HI8424; Hanna instruments).

2.2 Electrical characterization of the plasma

The electrical characterization of the DBDs was performed by measuring the energy dissipated per each voltage cycle. Instantaneous voltages (∆V) and charges (Q) were measured, respectively, with an HV probe (P6915A HV probe; Tektronix) and a 100-nF capacitor probe (Cmeas), connected in series to the ground electrode of the DBD and to a digital oscilloscope (TDS 2014 C; Tektronix). The energy dissipated per each voltage cycle was calculated by using the voltage-charge Lissajous method.[47] This energy was multiplied by the number of voltage cycles in one pulse and by the number of pulses in the treatment time, then divided by the area of the copper electrode (7.1 cm2) to calculate the energy dose delivered in each discharge. At least three measurements per DBD were performed per each set of conditions.

2.3 Detection of H2O2 and NO2− in PTWS

The detection of H2O2 and NO2− was carried with colorimetric assays of commercial test kits. A copper-phenantroline assay (cat. N. 118789; Merck Millipore) was used to detect H2O2, while the Griess assay (cat. N. 1.14776.001; Merck Millipore) was used to detect NO2− ions. Both assays were specifically validated in the case of cell culture media, as shown in our previous paper.[48] UV-Vis absorbance measurements were performed with a Cary 60 UV-Vis spectrophotometer (Agilent) in the spectral range 200–800 nm with 5 nm resolution. Disposable plastic cuvettes (cat. N. 223-9955; Bio-Rad) with a volume of 1 ml and 10 mm optical length were used. Further information about the validation of each assay in the specific case of cell culture media and PTWS can be found in our reference [48].

2.4 Statistical analysis and data processing

Statistically significant differences among data were computed with the software GraphPad Prism version 6.1. A one-way analysis of variance (ANOVA) was performed with a subsequent Turkey's multiple comparison test by assuming a normal distribution of the data. Statistical results of measurements were shown as the mean value ± standard deviation. All p values <0.01 were considered statistically significant.

2.5 Fourier-tranform infrared spectroscopy (FT-IR) on PTWS

FT-IR was used to inspect structural modifications of l-cysteine. Plasma treatments were performed on 2 ml l-cysteine solution (100 mg/L) in the same conditions used for the other PTWS (13 kV, 6 kHz, 25% DC, 0.5 slm air, 5 min); 500 µl of the treated solution were left drying overnight, under the hood, on clean shards of Si wafers (1.5 cm × 1.5 cm), at room temperature. FT-IR spectra (32 scans per analysis, 4 cm−1 resolution) were acquired in transmission mode with a Vertex 70 V Bruker spectrometer (Thermo Fisher Scientific). The spectrometer was evacuated to less than 150 Pa for 10 min before each acquisition. Spectra were elaborated with Bruker OPUS software.

3 RESULTS AND DISCUSSION 3.1 Electrical characterization of the DBDs

The DBD used in the experiments operates in a filamentary regime, with the formation of micro discharges during each voltage cycle, typical for a planar DBD.[49] The Lissajous figures obtained consist of closed parallelograms, as expected for capacitive systems.[50] By measuring the area of the Lissajous figure it is possible to calculate the energy dose of the plasma; as expected, this increases, with applied voltage (Figure 2a) and with treatment time (Figure 2b). The mean energy dose for an air DBD ignited at 13 kV for 180 s is 39  4 J/cm2. In Figure 2b, it is also shown that the energy dose is almost the same for O2, N2 and synthetic air.

Calculated plasma energy dose as a function (a) of applied voltage in a 180 s air DBD; and (b) of treatment time and gas feed in a 13 kV DBD

3.2 pH measurements of the liquids

The dissociation in the liquid of plasma-produced acid species like nitrous and nitric acids leads to nitrite and nitrate ions in PTWS, but also to the formation of H+ ions. This acidification could potentially affect the formation or the stability of other RONS, or the biological responses of living matter exposed to these liquids.[51] Therefore, it is important to monitor pH variations, above all when comparing different PTWS generated from buffered and non-buffered liquids. Only water is non-buffered, among the liquids under investigation, while PBS, DMEM and RPMI are buffered to neutral pH. Figure 3 shows the results of pH measurements of all tested liquids, before and after a mild (60 s, air plasma) and two harsh treatments (300 s, air and O2 DBDs).

pH measurements in water, PBS, DMEM and RPMI before and after a mild (60 s, air DBD) and two harsh (300 s, air and O2 DBD). All DBDs were ignited at 13.5 kV, 6 kHz, 25% DC, 0.5 slm, on 2 ml of each liquid, with 3 mm of treatment distance. DBD, dielectric barrier discharge; DC, duty cycle; DMEM, Dulbecco's modified Eagle medium; RPMI, Roswell Park Memorial Institute. One-way analysis of variance with Tukey's post-test: #p < 0.01 versus NT-water; *p < 0.01 versus NT-RPMI

As it is shown in Figure 3, in PBS and DMEM no significant alterations of pH were found, not even in the harsher condition. In the case of RPMI, a very small increase of pH was observed after plasma exposure. In the case of water, the milder treatment (60 s, air DBD) resulted in no pH variation, while prolonging the treatment to 300 s resulted in pH values around 4. No acidification was found in water exposed to DBDs fed with pure O2.

3.3 Analysis of NO2− in PTWS

The results of NO2− detection in water, PBS, DMEM and RPMI exposed to DBDs in different operating conditions are reported in Figure 4. The data demonstrate that the NO2− amount in PTWS strongly depends on the gas feed (Figure 4a), the treatment time (Figure 4b,c), the applied voltage (Figure 4d) and the volume of the treated liquid (Figure 4e). The values of NO2− concentration after the same plasma treatment were found different from one liquid to another. In particular, NO2− concentrations in PT-DMEM and PT-RPMI were found similar, but much higher than in PT-PBS and PT-water, as shown in each data series of Figure 4, especially when the O2/(O2 + N2) ratio is lower than 0.5. However, similar NO2− trends were found in each as a function of the plasma parameters. In all untreated liquids, the concentration of NO2− was below the detection limit.

NO2− concentrations in PT-water, -PBS, -DMEM and -RPMI as a function of (a) O2/O2 + N2 feed ratio; and treatment time in (b) air and (c) O2; (d) applied voltage; (e) liquid volume; (f) treatment distance.

Unless differently specified, all DBDs were ignited in air (0.5 slm) at 13 kV, 3 mm treatment distance, 2 ml of liquid, for 180 s. DBD, dielectric barrier discharge; DMEM, Dulbecco's modified Eagle medium; PT, plasma-treated; RPMI, Roswell Park Memorial Institute

The results here reported are coherent with the mechanism for NO2− production in PTWS reported in Figure 5. According to the literature, the main precursor of NO2− is gaseous HNO2, formed mostly in the plasma phase during •NO oxidation by •OH radicals (reaction 1 in Figure 5). When O2 and N2 are present in the feed, in fact, •NO is generated through the well known Zeldovich mechanism,[43] and it can also be detected in PT-water.[44] Moreover, further oxidation of •NO by O3 or O2 (reaction 2 and 3) can generate •NO2 in the plasma phase that reacts with water molecules, as vapour or in the liquid, to produce HNO2 as well (reaction 5 in Figure 5), that dissolves and dissociates in water to form NO2− ions, as reported in reaction 6 of Figure 5.

List of reactions in plasma and liquid phase involved in the formation of NO2− ions[31, 43, 44, 52, 53]

In agreement with the assumed mechanisms, hence, the production of NO2− ions appears strongly dependent on •NO and its derivatives in the plasma, whose formation in presence of N2 and O2 was assessed in our previous work.[44] In fact, the NO2− concentration depends on the N2 content in the feed (Figure 4a): N2-rich plasmas (O2/O2 + N2 < 0.4) maximize the amount of NO2−, while O2-rich plasmas minimize it. By working in O2 feed, in absence of N2, NO2− ions are practically absent in all PTWS (<20 µM) under investigation and no acidification of the liquids is found. This demonstrates that N2 is the limiting agent for the formation of nitrites. This trend is observed in all investigated liquids.

In parallel, an increase of the plasma energy dose, by raising the treatment time from 60 to 300 s (Figure 2d) or the applied voltage from 10 to 15 kV (Figure 2c), excites a greater number of N2 molecules in the plasma, which result in greater amounts of NO2− in the liquids (data in Figure 4b,d). Clearly, in absence of N2, increasing the energy dose generates only minimal NO2− amounts (<20 µM) in the liquids probably due to the residual air in the system (Figure 4c).

The above-mentioned considerations, though, do not account for the production of nitrite ions when no O2 is in the system (Figure 4a). When pure N2 is used in a closed DBD like ours, relevant O2 contaminations are quite unlikely. In case of a not well-sealed system, we would have observed relevant N2 contamination in O2 DBDs; with this feed, instead, we could not find NO2− in the liquids, attesting for the absence of air leaks. The only way to form nitrites or even NO would be to provide a source of oxygen in the plasma. As shown in Figure 4a, the optimal feed composition to form NO2- in the liquid is 95% N2 and 5% O2. This means that even small quantities of O2 promote the formation of NO2−. In pure N2, traces of O2 can be introduced by residual air, and by water evaporated from the liquid.

By fixing treatment times, applied voltage and feed composition, in principle, the amount of HNO2 produced in the plasma should be constant. Accordingly, the volume increase of the liquids from 2 to 5 ml dilutes HNO2 dissolved and, consequently, decreases the concentration of NO2− (Figure 4e). On the contrary, an increase in the treatment distance from 2 to 5 mm does not affect the amount of NO2− ions. Therefore, we can conclude that the NO2− formation in PTWS, in our conditions, is related to the presence of RNS, •NO or •NO2, in the plasma phase. Accordingly, the amount of NO2− in the liquids can be successfully optimized by properly tuning plasma feed gas composition, applied voltage and treatment time.

In spite of this, the concentration of NO2− ions in PT-cell culture media was found higher than in PT-PBS or PT-water, in each series of Figure 4, after the same plasma treatment. This difference should be related to the different chemical compositions of the liquids. The presence of inorganic salts (NaCl, KCl and others) is not responsible for the overproduction of NO2− in the culture media, as these species are contained in similar concentrations also in PBS. Instead, cell culture media contain high concentrations of organic molecules, which include amino acids, vitamins and carbohydrates. Our hypothesis, in fact, is that these molecules can be involved in nitrosation reactions with transient RNS like NO, NO2 or N2O3 to form nitro-organic molecules that could release additional NO2− ions.

Nitrosation reactions generally involve electrophilic species like NO+ or NO, capable of reacting with electron donors through the nitrogen atom.[54] In cell culture media with neutral pH, however, the possibility to find NO or NO+ is negligible because they persist only in strong acid conditions.[54] Nonetheless, it is possible that the nitrosation of the biomolecules involves NO+ donor species like N2O3, which can be formed by plasma-generated NO and NO2, and dissolves in aqueous media (reaction 4, in Figure 5),[23] with the consequent release of NO2− ions. Thiol groups of sulfur-containing amino acids[44, 53-56] like cysteine and methionine, or the amino groups[54, 57] on side chains of many amino acids may be targeted by N2O3, according to reactions 7 and 8 in Figure 5, effectively releasing nitrite ions in the media.

Demonstrating this chemical mechanism, however, would require specific kinetics experiments. We have previously investigated the production of •NO in PT-water and in PT-DMEM and confirmed that, while •NO was detected in PT-water, it was absent in PT-DMEM due to the scavenging action of some organic compounds (d-glucose in prevalence) and to the buffered pH, far from the acid conditions needed for •NO formation in liquids.[44] Moreover, several studies report thiols nitration in l-cysteine and l-methionine to form S-nitroso-organic compounds in plasma-treated liquids.[44, 53, 55, 56]

For this reason, we have used FT-IR spectroscopy to record the chemical fingerprint of an l-cysteine solution after an air DBD to investigate the structural modification of the biomolecule. The treatment was performed on 2 ml of cysteine solution in water (100 mg/L), not acidified with HCl, in the same conditions used for the other PTWS (13 kV, 6 kHz, 25% DC, 0.5 slm air, 5 min). 500 μl of the treated solution were left drying overnight, under the hood, on Si shards (1.5 × 1.5 cm), at room temperature. The FT-IR spectra of untreated (NT-Cys) and plasma-treated cysteine solution (PT-Cys) are shown in Figure 6.

FT-IR spectra of untreated (NT-cys, in black), and plasma-treated cysteine solution (PT-cys, in red)

Structural changes in PT-Cys include (i) oxidation of -SH groups to -S═O (νSO at 1040 cm−1); (ii) formation of -OH groups (νOH, 3200–3500 cm−1); (iii) formation of -COOH groups (ν asymCOOH, 1610 cm−1); (iv) nitrosation of -SH group (resonance structure of S-nitrosothiol group, νN═O, 1487 cm−1). It is possible that oxidation and nitrosation of -SH group may occur at the same time, on different cysteine molecules in the solution. In this way, the nitrosation of the thiol groups may effectively be a possible secondary route for NO2− production in culture media. Although FT-IR spectra support this hypothesis this last aspect needs to be fully elucidated.

3.4 Analysis of H2O2 in PTWS

The results of H2O2 detection in DMEM, RPMI, PBS and water exposed to different DBDs are reported in Figure 7. Unlike the production of NO2− ions, the data indicate that the production of H2O2 depends very strongly on the type of liquid treated rather than on the plasma conditions varied. Different trends are observed between the two cell culture media and water or PBS.

H2O2 concentrations in PT-water, -PBS, -DMEM and -RPMI as a function of (a) O2/O2 + N2 feed ratio; treatment time in (b) air and (c) O2; (d) applied voltage; (e) liquid volume; (f) treatment distance.

Unless differently specified, all DBDs were ignited in the air (0.5 slm) at 13 kV, 3 mm treatment distance, far from 2 ml of liquid, for 180 s. DBD, dielectric barrier discharge; DMEM, Dulbecco's modified Eagle medium; PT, plasma-treated; RPMI, Roswell Park Memorial Institute

In particular, it is observed that in our conditions the formation of H2O2 in water and in PBS is inhibited (<10 µM, the limit of detection of the colorimetric assay[48]) in all conditions, despite the wide range of parameters tested (Figure 7). On the contrary, in the two culture media, the production of H2O2 is observed, with more scattered trends as a function of the individual plasma parameters with respect to NO2−. Moreover, the concentration of H2O2 in PT-DMEM was found higher than in PT-RPMI after the same treatments, in all cases, as it clearly emerges in Figure 7. In both culture media, an increase in the H2O2 amount is observed by increasing the O2/O2 + N2 ratio (Figure 7a), the treatment time (Figure 7b,c) and the applied voltage (Figure 7d). On the contrary, it remains almost constant by increasing liquid volume (Figure 7e) and treatment distance (Figure 7f).

Indeed, the results here reported strongly challenge the mechanism of H2O2 production in PTWS assumed so far in the literature, schematically reported in Figure 8. It is generally mentioned that the main precursor of H2O2 in liquid is the •OH radical,[26, 27, 42, 58-60] which is formed in the plasma from H2O molecules by dissociation through electron collisions (reaction 1), photodissociation (reaction 2), or its reaction with •O atoms (reaction 3).[52] The formation of H2O2 is mostly attributed to •OH recombination in the plasma (reaction 8, in presence of a collision partner M) or in the aqueous phase (reaction 9), where •OH is highly soluble.[52] Moreover, also the hydroperoxyl radical HO2• (reaction 4) can react with •H (reaction 5) or with •O (reaction 6) atoms to restore •OH, or it can recombine to form H2O2 (reaction 7). The occurrence of such a mechanism is clearly related to the presence of water vapour molecules in the plasma phase to produce •OH or HO2• radicals.

List of reactions in plasma and liquid phase involved in the formation of H2O2 in treated liquids[27, 42, 52, 58, 59]

Despite this, however, we could not find H2O2 in PT-water or PT-PBS, while it was detected in high concentration in both PT-culture media. Indeed, the organic molecules and the buffered pH in these liquids should quench the oxidative capacity of RONS and reduce their concentration.[44] Many groups, though, have recently reported that organic compounds increase rather than decrease the production of RONS in PTWS. For example, Privat-Maldonando et al.[46] reported that the lack of bactericidal activity in plasma-treated solutions with high organic content, like DMEM and other culture media, is correlated with the decrease of transient RONS, such as •OH, O3 or •O and with the increase of H2O2 and NO2− (with lower bactericidal effect).[46] Similarly, also Ranieri et al.[59] have found long-lived RONS more concentrated in PT-RPMI than in PT-water after the same treatment, and concluded that reactions between medium components and plasma-produced species may have led to the overproduction of long-lived RONS in RPMI.[