1. Introduction
Non-thermal plasmas (NTP) intensified appear as a decontamination technique and tend into the focus of research [
1,
2]. The energy of NTPs mainly houses in the free electrons, which results in a partial ionization of the gas and low gas temperatures [
3,
4]. However, a low gas temperature opens the possibility for plasma techniques to gain a foothold in new areas. NTP treatments can be roughly divided into direct (where a plasma source is directly acting on the target) and indirect treatments (where a plasma-treated media is used to treat the target) [
5]. Plasma-treated water (PTW) is such an indirect method. It houses an anti-microbial activity, which is mediated by chemical reactions that were triggered during plasma-processed air (PPA) generation.
PTW houses a cocktail of active chemical compounds, which have been generated during its production. However, plasma-treated water (PTW) is such an indirect method. For the generation of PTW, different types of plasma sources are used, e.g., DBD, and plasma torches and jets. In addition, there is a huge variety of process schemes for the interaction of plasma with water [
6]. In some applications, the water surface acts as an electrode; in others, the plasma has no direct contact with the water, and only the working gas, modified by the plasma interaction, is brought into contact with the water. In the present work is focused on PTW with high amounts of RNS. Therefore, microwave plasma sources were developed because of their ability to generate plasma in the temperature regime of 3.500 K, which is known to be very efficient in NOx generation. Due to their relatively high temperatures, microwave-induced plasmas push the reactions that are initiated in PTW into the direction of reactive nitrogen species. Therefore, PPA generation is often equated with the Birkeland–Eyde process [
7]. However, since the Haber–Bosch process was invented, the Birkeland–Eyde process never played a dominant role. However, in the scope of a novel decontamination technique, nitrous gasses appear in a new light.
Whereas long-living compounds such as HNO
3 or H
2O
2 are easy to analyze, many authors postulate short-living components of PTW and address them as active anti-microbial components. Additionally, several studies suggest a higher impact on the pH value of the solution than its ingredients. However, the composition of the PTW is not only governed by way of its production; the specific geometry of the device, the production temperature, or the type of plasma generation also influences the cocktail. PTW was able to demonstrate its anti-microbial effect on suspension cells and biofilms. Several authors describe the positive impact of PTW on foodstuffs [
8,
9,
10,
11]. For instance, Schnabel et al. (2021) describe the usage of PTW in a washing line of leafy greens [
12]. However, more than just produce is anti-microbially processed with PTW. PTW is feasible for the cleaning of surfaces with a high anti-microbial decontamination rate [
13,
14]. It can be used as a bath for processing tools or as a sanitizer for wider, unspecific purposes [
15]. Additionally, plasma appears in agriculture to increase agricultural production. For instance, Brust et al. (2020) used a dielectric barrier discharge system for the short-term treatment of cereal seeds [
16]. The authors describe improved germination of the seeds. Plasma medicine is also an emerging field of plasma technology [
17]. The anti-microbial potential opens new ways to combat pathogens such as bacteria or viruses. For instance, Guo et al. (2018) use PTW for efficient antiviral inactivation [
18]. PTW also combats bacteria that gain antibiotic resistance and has the feasibility to open new perspectives and treatment schemes [
19].
A growing number of researchers characterize PTW based on concentration measurements, pH [
20], electron paramagnetic resonance (EPR) [
21], or oxidative reaction potential [
22]. At the same time, all the above-described methods gather information that describes the solution without being able to deduce a procedure for the mixing of PTW. Against that background, ionic chromatography and chronoamperometry appeared as ideal methods for the quantitative characterization of PTW. It is the aim of this study to generate a PTW analog (anPTW) from traceable compounds found in freshly produced PTW. The composition of the solutions is based on quantitative PTW analysis. The solutions, which are the anPTW, an HNO
3-, an HNO
2-, and an H
2O
2-solution, should be as potent in their anti-microbial activity as PTW. The decisions are based on a combination of three biological assays. This study presents an evaluation method where it compares every solution with PTW. The outcomes of the tests can be subdivided in six categories, which allows a classification of the solutions.
3. Results
3.1. PTW and Its Compounds
The generation of plasma-processed water (PTW) started with compressed air carried over a microwave-driven plasma torch. Subsequently, the air was processed in a manner dependent on the used plasma source. When microwave sources were used, predominantly reactive nitrogen species and, to a lesser extent, reactive oxygen species were produced. Subsequently, the PPA has carried over pipes, where it cools down. Consequently, PPA processed via a microwave source houses nitroxides pushed into a metastable state. Traceable, long-living compounds of PPA were determined via Fourier Transform Infrared spectroscopy (FTIR).
Figure 2a summarizes the findings of the measurements. NO
2 was detected in the range of 1540–1660 cm
−1, with the second order of energy absorption in the range of 2840–2938 cm
−1. NO was observed in the range of 1790–1950 cm
−1, while N
2O
5 was detected in the 700–779 cm
−1, 1220–1292 cm
−1, and 1668–1791 cm
−1 ranges. Due to the existence of water under high humidity operation conditions, a small amount of nitric acid was produced and observed in the spectra of PPA. As follows, NO
2 takes up more than 70% of the quantity of the key reactive species; it is, therefore, marked as an indicator for deciding optimal process parameters.
PTW is a chemical cocktail of compounds that originates from chemical reactions started by the action of a plasma source. As is observed for PPA, the compounds found in the processed water originate from the metastable and long-lived RONS. Those traceable and long-living species were determined via ion chromatography and potentiometric measurements.
Table 1 summarizes the findings of the analytical measurements of PTW.
Figure 2b summarizes the postulated and/or literature findings of possible reactions in PTW. The most prominent compound found in PTW is HNO
3 (33mol/L). Additionally, we observe a concentration of 1.17 mol/L of HNO
2. Both compounds can be generated from NO
2 of PPA. The reactive oxygen species H
2O
2 was also found in freshly generated PTW (21.1 mol/L). All compounds monitored in PTW have been solved in water.
3.2. Anti-Microbial Activity of PTW against Suspension Cells of Pseudomonas fluorescence
Figure 3 shows the outcomes for the treatment of suspension cells with PTW, anPTW, HNO
3, HNO
2, and H
2O
2.
Figure 3 upper part shows all data of all solutions merged in one sketch. The graph represents the magnitude of all values relative to each other.
Figure 3 lower part reflects the comparison of the anti-microbial effect of PTW with the effect of solutions of its ingredients. Additionally, the impact of PTW on suspension cells is compared with the impact of its analog (anPTW) made from the traceable compounds found in IC- and potentiometric measurements.
For suspended P. fluorescence cells in a proliferation assay (CFU), the action of PTW after 1 min treatment time compared with the impact of anPTW after the same treatment time show no statistically meaningful difference for their RFs (PTW vs. anPTW, RF, 1 min: 2.61 ± 0.52 vs. 2.14 ± 1.31, ANOVA: p < 0.05). Nevertheless, both treatments reveal a statistically meaningful decrease in the cell numbers in that assay (ANOVA: p < 0.05). The observation continues for a 3 min and 5 min treatment, where no further meaningful changes (ANOVA: p > 0.05) appear for both kinds of treatments (PTW vs. anPTW, RF, 3 min: 2.91 ± 0.62 vs. 3.14 ±1.11; 5 min: 3.04 ± 0.75 vs. 3.67 ± 0.70). The XTT assay follows that tendency, and after a drop (PTW and anPTW, MANOVA: p < 0.05) of the metabolic activity of the cells for both treatment solutions, no further statistically meaningful changes (PTW and anPTW, ANOVA: p > 0.05) have been determined (PTW vs. anPTW, 3 min: 0.32 ± 0.86, vs. 0.19 ± 1.39, 5 min: 0.45 ± 0.21 vs. 0.22 ± 0.09). The L/D assay for PTW underpins the interpretation of the CFU- and XTT assay. A statistically meaningful drop (PTW, ANOVA: p < 0.05) of the membrane integrity is observed after treatments of 1 min, which show no further meaningful (PTW, ANOVA, p < 0.05) drops for longer treatments. anPTW treatments show no statistically meaningful impact on the membrane integrity (anPTW, ANOVA, p < 0.05). A treatment with anPTW pushes the cells into VBNC, whereas a PTW treatment appears successful.
Treatment of suspended cells with HNO3 appears ineffective when compared with that of a PTW treatment. First, no statistically meaningful (HNO3, ANOVA, p < 0.05) changes in the cell count are observed after a 1 min treatment (HNO3, RF, 1 min: −0.04 ± 0.06); a tendency that continues after a 3 min and 5 min treatment time (HNO3, RF, 3 min: 0.06 ± 0.07, 5 min: −0.04 ± 0.08). There are no statistically meaningful changes in cell number (HNO3, ANOVA: p > 0.05). The L/D assay completes the picture and now shows statistically meaningful changes (HNO3, ANOVA: p > 0.05) in the relation between living and dead cells (HNO3, all values are relative to their reference, 1 min: 1.33 ± 0.74, 3 min: 1.27 ± 0.81, 5 min: 1.29 ±0.79). All PTW-induced changes in the cell number, their metabolic activity, and their membrane activity appear statistically meaningful when compared to the ones obtained after an HNO3 treatment. An HNO3 treatment was judged as an ineffective failure.
An HNO2 treatment of suspended cells appears effective but to a lesser extent. At the same time, the CFU count is not statistically meaningful or different from that of a PTW treatment (PTW vs. HNO2, ANOVA: p > 0.05). The means of the cell numbers are quite different (PTW. vs. HNO2, RF, 1 min: 2.61 ± 0.52 vs. 1.30 ± 1.11), but huge error bars suffer the mean of the CFU number. Both treatments show no further statistically meaningful (PTW and HNO2, ANOVA, p < 0.05) impact for longer treatment times (HNO2, RF, 3 min: 1.09 ± 0.95, 5 min: 0.66 ± 1.41). The XTT assay follows that tendency, and no meaningful changes (HNO2, ANOVA: p < 0.05) were determined throughout the whole experimental series (HNO2, all values are relative to its reference, 1 min: 0.51 ± 0.18, 3 min: 0.44 ± 0.21, 5 min: 0.45 ± 0.21). The L/D assay also follows the tendency, and again, there were no statistically meaningful changes (HNO2, ANOVA: p < 0.05) of the cell integrity due to an observed HNO2 treatment (HNO2, all values are relative to their reference, 1 min: 0.12 ± 0.83, 3 min: 0.19 ± 0.72, 5 min: 0.15 ± 0.68). An HNO2 treatment seems to detach dead cells, which leads to a drop in the cell number but to rather invariable values for the metabolic activity or the integrity of the cell membrane of P. fluorescence.
The H2O2 treatment of suspended cells has a CFU drop after a 1 min treatment of a comparable magnitude of a PTW treatment (PTW vs. H2O2, RF, 1 min: 2.61 ± 0.52 vs. 2.72 ± 0.56) and no statistically meaningful difference (PTW vs. H2O2, ANOVA, p < 0.05) could be detected. Usually, longer treatment times lead to a higher RF. Contrarily, an H2O2 treatment reveals decreased RFs for treatment times of 3 min and 5 min (H2O2, RF, 3 min: 0.63 ± 0.61, 5 min: 0.98 ± 1.03). The changes appear to be statistically meaningful (H2O2, ANOVA: p < 0.05). Unlike the proliferation assay, the XTT assay shows no meaningful (H2O2, ANOVA: p > 0.05) decreased metabolic rate due to an H2O2 treatment relative to their references. The L/D assay depicts no statistically meaningful (H2O2, ANOVA: p > 0.05) drop in the cell integrity for all treatments (H2O2, all values relative to their references, 1 min: 0.83 ± 0.22, 3 min: 0.72 ± 0.174, 5 min: 0.68 ± 0.19). The result is interpreted as the detachment of dead cells.
3.3. Anti-Microbial Activity of PTW against Biofilms of Pseudomonas fluorescence
Figure 4 shows the outcomes for the treatment of biofilms of
P. fluorescence with PTW, anPTW, HNO
3, HNO
2, and H
2O
2.
Figure 4 upper part shows all data of all solutions merged in one sketch. The graph represents the magnitude of all values relative to each other.
Figure 4 lower part reflects the comparison of the anti-microbial effect against biofilms of PTW with the effect of solutions of its ingredients. Additionally, the impact of PTW is compared with the impact of its analog (anPTW) made from the traceable compounds found in IC- and potentiometric measurements.
For biofilms of P. fluorescence in a proliferation assay (CFU), the action of PTW after 1 min treatment time compared with the impact of anPTW after the same treatment time showing no statistically meaningful difference (MANOVA, p > 0.05) for their RFs (PTW vs. anPTW, 1 min: 3.97 ± 0.51 vs. 5.74 ± 2.79). The observation continues for a 3 min treatment. As opposed to this, a statically meaningful difference (p < 0.05) is observed for a 5 min treatment (PTW vs. anPTW, 5 min: 3.73 ± 0.62vs. 8.08 ± 1.77). The XTT assay draws a picture and vice versa. Both solutions show a drop in their metabolic activity when compared with their reference (PTW vs. anPTW, 1 min: 0.26 ± 0.33 vs. 0.62 ± 0.22). For both solutions, these changes are statistically meaningful (MANOVA, p < 0.05). Based on the drop (ANOVA, p < 0.05) of the metabolic activity after a 3 min PTW treatment (PTW, 3 min: 0.19 ± 0.18), the PTW treatment is interpreted to have a stronger adverse effect on suspended cells and biofilms than anPTW. Both solutions show no further statistically meaningful (MANOVA, p > 0.05) drop in the metabolic activity for longer treatments. The L/D assay for PTW underpins the interpretation of the CFU and XTT assay. A statistically meaningful drop of the membrane integrity is observed after a treatment of 1 min (PTW, 1 min: 0.29 ± 0.14), which shows no further meaningful (p < 0.05) drops for longer treatments. After a 1 min anPTW treatment of suspended P. fluorescence cells and biofilms, no statistically meaningful change of the cell integrity is observed (anPTW, 1 min: 0.84 ± 0.16). However, the drop of the membrane integrity became meaningful after an anPTW treatment of 5 min (anPTW, 5 min: 0.62 ± 0.11). A treatment with anPTW may be successful after 5 min, but at least a 1 min treatment pushes the cells into VBNC.
The comparison of the PTW treatment vs. HNO3 is unambiguous. The 1 min PTW treatment shows an RF of 3.97 ± 0.51 vs. an RF of 0.21 ± 0.37 after an HNO3 treatment. For both treatment solutions, no statistically meaningful changes have been observed for longer treatments (PTW vs. HNO3, 3 min: 4.05 ± 0.65 vs. 0.25 ± 0.29, 5 min: 3.73 ± 0.62 vs. 0.35 ± 0.34, ANOVA: p < 0.05). The trend continues for the XTT assay where a statistically meaningful drop of the metabolic activity is observed after a 1 min PTW treatment (PTW relative to its reference, 1 min: 0.26 ± 0.33, ANOVA: p < 0.05). No statistically meaningful further changes for longer treatments were revealed (PTW relative to its reference, 3 min: 0.19 ± 0.18, 5 min: 0.34 ± 0.21, ANOVA: p > 0.05, ANOVA: p > 0.05). An HNO3 treatment of biofilms shows no statistically meaningful (p > 0.05 for ANOVA) changes at any treatment times (HNO3 relative to its reference, 1 min: 0.91 ± 0.21, 3 min: 0.88 ± 0.23, 5 min: 1.01 ± 0.07). The L/D assay makes the same statement. While a meaningful (ANOVA: p < 0.05) drop in the cell integrity of P. fluorescence biofilms is revealed after a 1 min PTW treatment (PTW relative to its reference, 1 min: 0.29 ± 0.15), an HNO3 treatment showed no meaningful drop (HNO3 relative to its reference, 1 min: 0.86 ± 0.16, ANOVA: p > 0.05). The tendency is conserved for longer treatments (PTW vs. HNO3, all values relative to their references, 3 min: 0.26 ± 0.14 vs. 1.09 ± 0.29, 5 min: 0.25 ± 0.15 vs. 0.93 ± 0.27) and no statistically meaningful changes have been observed (PTW, ANOVA: p > 0.05; HNO3, ANOVA: p > 0.05). The successful PTW treatment opposes a failure in terms of the anti-microbial action of the HNO3 treatment.
The comparison of the anti-microbial active PTW treatment with those of an HNO2 treatment reveals statically meaningful RF for both treatments (PTW vs. HNO2, 1 min: 3.97 ± 0.51 vs. 5.65 ± 2.71) but no statistical difference (ANOVA: p > 0.05). Nevertheless, they become meaningful after a 3 min and 5 min treatment (PTW vs. HNO2, all values relative to their references, 3 min: 4.05 ± 0.65 vs. 8.88 ± 3.97, 5 min: 3.73 ± 0.62 vs. 9.15 ± 1.63). The XTT assay shows no meaningful changes for an HNO2 treatment for all treatment times (HNO2, ANOVA: p > 0.05). The PTW treatment has a meaningful (see above) drop in metabolic activity, which shows no further impact for longer treatment times. The comparison of a 5 min PTW treatment with an HNO2 treatment of the same length shows no meaningful difference in their outcomes (ANOVA: p < 0.05). However, since a PTW treatment shows a statically meaningful change for a treatment up to 5 min (ANOVA: p < 0.05), we interpret these findings as decreased cell integrity due to a PTW treatment and rather invariable values or only light changes for an HNO2 treatment (HNO2 relative to its reference, 3 min: 0.87 ± 0.31, 5 min: 0.69 ± 0.37). The HNO2 treatment with its high RFs in the proliferation assay is interpreted as a detachment of dead cells with a possible small number of living or VBNC cells of P. fluorescence.
An H2O2 treatment reveals, when compared with a PTW treatment of the same manner, meaningful changes (ANOVA: p < 0.05) in their cell count (H2O2, RF, 1 min: 5.12 ± 1.96) after a treatment time of 1 min. No further meaningful (H2O2, ANOVA: p > 0.05) changes are observed for longer treatment times (H2O2, RF, 3 min: 5.12 ± 2.97, 5 min: 4.13 ± 2.28). Both the XTT and the L/D assays show no statistically meaningful (ANOVA: p > 0.05) decrease for the metabolic activity of the cells (H2O2 relative to their references, XTT, 1 min: 0.74 ± 0.26, 3 min: 0.92 ± 0.21, 5 min: 1.01 ± 0.0) or an adverse impact of an H2O2 treatment on the cell integrity (H2O2 relative to their references, L/D, 1 min: 0.86 ± 0.21, 3 min: 0.91 ± 0.15, 5 min: 1.01 ± 0.29), respectively. The results have been interpreted as a VBNC state of an H2O2 treatment opposing the successful PTW treatment.
4. Discussion
PTW is produced by means of PPA. That is, ordinary compressed air is processed with an ignited plasma torch, which is subsequently passed into distilled water. This step initiates chemical reactions that are at the beginning of every PTW action. Generally, the choice of a certain plasma source brings the opportunity to push chemical reactions on the desired path. For non-thermal plasmas (NTP), the main conversion rates of certain chemical reactions are either on the side of reactive oxygen or on the side of reactive nitrogen compounds [
24,
32]. In the case of microwave-generated plasma, the main conversion rates for atmospheric compounds are on the oxidation of nitrogen [
2]. A process that resembles a Birkeland–Eyde process, which is basically a plasma process using an electric arc [
7]. The process has been used for the production of nitrous gases as a starting product for several products of the chemical industry, such as ammonia, fertilizer, or explosives [
7,
33].
Directly after the processing of the compressed air, the freshly produced PPA was studied by FTIR measurements, and the radical NO
2 was the main component that was detectable. Additionally, the gaseous NO has been observed. In addition to that finding, the spectra reveal the presence of the solid N
2O
5. However, N
2O
5 has a sublimation point that is very close to room temperature (32 °C). Against that background, it is likely to detect sublimated N
2O
5 in the gaseous PPA. It is surprising that no N
2O
4, the dimeric of NO
2, was observed in the spectra. Most likely, the concentration is too low. The IR-active chemical compounds have been the expected outcome for such a measurement [
34]. The measurements underpin the statement of the similarity of the Birkeland–Eyde process and the presented PPA generation via a microwave-driven plasma source [
7].
Subsequently, the PPA is passed into distilled water, and at the liquid–gas interface, gaseous reactive nitrogen species enter the liquid [
24]. Consequently, microwave-generated PTW houses a set of nitrous compounds that provide anti-microbial efficacy [
35]. The study aimed to answer whether it is possible to mix anPTW from the results of analytical measurements such as IC, pH measurements, or electrochemical characterizations of the PTW. As a first step, IC measurements revealed a high content of nitrous and nitric acid. Nitrous acid appeared as the main component of the PTW solution and resulted from the reaction of nitrogen dioxide, which entered the solution as a part of PPA, with water [
4]. Additionally, NO
2, in a reaction with H
20, is also a source for HNO
3. However, we observe different concentration ratios for HNO
2 and HNO
3, which can be greater or less than 1 depending on the plasma source used. Further PTW reactions, which are initiated by passing PPA, solely postulate the presence of peroxynitric acid (HNO
4) or the conjugated base peroxynitrite (OONO
−), which is a short-living component, and its detection appears challenging. In the literature, it is also described as a reservoir for NO
2 through a reversible radical reaction [
36]. However, the characterization of metastable (in addition to N
2O) nitrous gases is not a new field of research, and many scientists gained knowledge through their research. Since nitrous compounds are ubiquitous, researchers from many fields enter the scene [
36,
37,
38]. Additionally, cracked and oxidized nitrous compounds are pushed into a metastable state due to cooling down after the process. As early as the generation of PPA, the relative composition of PTW strongly depends on the composition of PPA, which again depends on the generation temperature and the subsequent cooling down influenced by the source set up [
32]. Generally, the composition of PTW depends on many factors such as the type of the source, excitation properties, or the gas flow rate [
39]. Against that background, it is advantageous to consider the total content of all nitrogen compounds and not individual components [
40,
41].
Although only H
2O
2 has been detected, PTW also houses reactive oxygen species [
42]. Due to a higher power density of a microwave-induced plasma, the process is performed at higher temperatures as reached by radiofrequency plasmas. Against that background, it is rather likely to find a plasma gas containing a higher proportion of reactive nitrogen species than reactive oxygen species [
9]; PTW generates the reactive components of PPA, which is predominantly NO and NO
2. Theses compounds react to HNO
3 and HNO
2 when introduced into PTW. Beside those reactions, H
2O
2 have been also found in PTW: However, the H
2O
2 concentration was higher than the HNO
2 concentration but much lower than the HNO
3 concentration. It is an expected result for PTWs produced with a microwave-induced plasma device.
It was the aim of the study to answer the question of whether it is possible to mix an analog PTW out of commercially available chemicals. More importantly, are these solutions as anti-microbial efficient as freshly produced PTW? Against that background, a series of assays (CFU, XTT, and L/D) have been performed on suspension cells and biofilms of P. fluorescence. The assays monitor cell proliferation (CFU), metabolic cell activity (XTT), and cell integrity. Since all tests have been performed in a dynamic time frame, the results of the assays are suffered by high standard deviations. To answer the question posed at the beginning of the chapter, we statistically evaluated the results in terms of a direct comparison of PTW with every traceable ingredient found in freshly produced PTW. The anti-microbial success of treatment with one of the solutions is judged on an arrow system described earlier.
An anti-microbial treatment against suspended P. fluorescence and mono-species biofilms with PTW appears successful. We see a drop in the proliferation followed by a decreased metabolic activity and results from the L/D assay lower than one, which is, in combination with the two other tests, a drop in the number of living cells. When such treatment is compared with anPTW, the anPTW treatment appears as effective as a PTW treatment since we see a decreased proliferation. However, when all three assays are taken into account, no decreased cell integrity is observed. Against that background, it is likely to assume more cells have entered VBNC. A decreased proliferation and metabolic activity speak for the mortality of these cells. Contrarily, for killed suspension cells, an L/D value lower than one can be expected. A test–outcome combination very likely speaks for cells entering the VBNC state.
A comparison with HNO
3, which is a concentrated compound found in PTW, did not show any anti-microbial efficiency for both suspension cells and biofilms. No meaningful changes are observed for the whole assay series. The result is interpreted as an unsuccessful treatment. The result is rather confusing since HNO
3 is frequently used in CIP procedures to combat adverse biofilms [
43]. However, since CIP procedures work with different solutions from the alkaline to the acidic end of the spectra, it is rather likely that a combination of active agents leads to success. In addition to that, although the concentration of HNO
3 is relatively high in PTW, higher concentrations may have a more pronounced effect on microorganisms.
A comparison of HNO
2 action on suspended cells and biofilms of
P. fluorescence reveals a process that pushes suspended cells into VBNC and detaches dead cells from the biofilm. The arguments for interpreting the outcomes of the combined tests as VBNC do not differ from the ones given before. The detachment of cells follows from a decreased proliferation framed by no statistically meaningful changes in the metabolic rates and the integrity of the cells. Compared with a PTW treatment, an HNO
2 treatment appears less effective. Especially if only the CFU is taken into account, the treatment is less effective against suspension cells but more effective when biofilms are combated. The XTT assay of HNO
2-treated cells and biofilms varies for all treatment times and seems to be decreased, but huge error bars suffer it. Here, no statistically meaningful changes could be detected. Overall, an HNO
2 treatment pushes suspended cells into VBNC and detaches biofilms. Nevertheless, VBNC can be interpreted as a cellular stress reaction induced by stress such as starvation, osmotic or oxidative stress, or adverse nutrition [
44,
45,
46,
47]. Stress is always a matter of an offensive descent on bacterial cells. Is it only one or a few components that mediate the stress, or is a cell very likely to be in the position to defend itself? Nevertheless, if, as is the case for PTW, the solutions have a potpourri of different anti-microbial components, it becomes likely that a cell lacks sufficient defense mechanisms, which is a stress that treatment with HNO
2 could well provide. Nevertheless, it does not provide the anti-microbial efficiency of PTW, which rather kills the cells instead of pushing them into VBNC. For biofilms, a cell detachment could be as effective as the killing of the cells, but it houses the danger of cross-contamination.
Treatment of suspended cells of
P. fluorescence with H
2O
2 equals an H
2O
2 treatment of biofilms of
P. fluorescence. Both H
2O
2 treatments push suspended cells and biofilms into the VBNC state. For both manifestations of
P. fluorescence, slight changes in the L/D assay are visible. Albeit they are not statistically meaningful, the changes depict a contrary effect. Whereas a decrease in cell integrity with longer treatment times is observed for suspension cells, it seems to increase with longer treatment times for biofilms. H
2O
2 treatments seem to be successful but show no meaningful results. The H
2O
2 mechanism of action is interpreted as cellular stress, which might be opposed by cellular protection mechanisms [
48,
49]. An H
2O
2 treatment has been interpreted as less effective than a PTW treatment.
If all outcomes of the treatments against suspended cells and biofilms of
P. fluorescence are added up, PTW appears as the most powerful anti-microbial solution among the tested ones. Either a solution of a diluted component found in PTW or anPTW could fully explain the anti-microbial efficiency and its magnitude. It is obvious that PTW houses a complex chemistry, which is not explained by the determination of traceable long-living species. Several authors postulate various short-living species, such as peroxynitrite or even complex water structures in PTW [
41,
50,
51]. Deeper future investigations need to explain the excellent anti-microbial potential of PTW. Additionally, Niquet et al. (2017) chemically compare DBD and microwave-generated PTW and the biological effect of the compounds found in these solutions [
24]. These authors demonstrate how the concentrations of the compounds are dependent on the type of generation. Against that background, our study demonstrates for the first time that PTW has better anti-microbial activity than a mixture of its detectable components. The exact mechanism of PTW action needs to be clarified in future studies.