Non-thermal atmospheric pressure plasmas are widely used in biomedical research and clinical applications. Such plasmas generate a variety of reactive oxygen and nitrogen species upon interaction with ambient surroundings. These species further interact with a biological substrate and are responsible for the biomedical effects of plasma. Liquid water is an essential part of any biological systems. Some of the most reactive species induced by plasma in aqueous media are radicals and atoms. Hence, the presence of certain chemical components in a plasma ‘cocktail’ presents an important task for both understanding and further development of plasma systems with specific purposes. In this chapter, we discuss various methods of detection of the plasma-generated short-lived reactive species. We dissert various plasma-induced radicals and atoms (•OH, O2•−/•OOH, •NO, O), together with non-radical short-lived species (−OONO, O3, 1O2). Electron paramagnetic resonance (EPR) is the most direct method of radical detection in water-based media. Special attention is paid to the limitations of the detection methods, with an emphasis on spin trapping used in EPR analysis.
- plasma-liquid systems
- reactive species
- free radicals
- spin trapping
- electron paramagnetic resonance
Low-temperature, or ‘cold’ atmospheric pressure plasmas (CAPs) are gaining increasing attention in diverse fundamental and applied scientific activities . The research on the industrial applications of cold plasma includes its use in plasma-assisted catalysis and thin film deposition [1, 2, 3], wastewater treatment [4, 5], photoresist removal , pre-treatment of polymeric solutions for the production of enhanced nano-fibres , etc.
Biomedical applications are the most burgeoning field of CAP research . Cold plasma is used to modify or produce surfaces with high bacterial resistance, an important property in a clinical setting [8, 9]. Other applications include various sterilisation processes , deactivation of bacteria and viruses [11, 12], wound healing , and the emerging CAPs cancer treatment [8, 13]. The effects of CAPs on biological substrates are largely defined by the reactive oxygen and nitrogen species (RONS) such as •OH, O2•−/•OOH, O3, 1O2, O, •NO, H2O2, ONOO−, NO2− . These species are formed either in the plasma itself, or upon its interaction with surrounding air [14, 15]. Water is an essential component of every biological system. Thus, information on both the composition of the mixture of RONS created by plasma, and their interactions with aqueous media is extremely important for tailoring-desired plasma effects .
Chemical modelling coupled with various analytical techniques (optical spectroscopy methods, mass spectrometry, etc.) is used to assess the composition of the gas phase plasma [1, 14, 17]. Recent works in computational chemistry have addressed the interaction of gas phase RONS with and within aqueous media [18, 19]. However, monitoring of the reactive species in liquid is paramount for benchmarking of the models. Most importantly, it provides experimental, and hence the most direct information on RONS present in the liquid.
Two main paths are used for CAP utilisation in biomedical research and applications: first, pre-treatment of a relevant medium with further application to the biological target [20, 21] and second, direct application of plasma treatment to cells in aqueous media , or ‘dry’ cells  (We note that generally all cells are grown in culture medium and/or washed prior to plasma exposure; thus, the ‘dry’ cell surface is never devoid of water, even less, so is a tissue in clinical applications ). In the first scenario, the effects of plasma are attributed to long-lived chemical species, which can remain in solution after plasma treatment, such as H2O2, NO2−, and NO3− [20, 25]. These long-lived species are usually detected using a variety of analytical techniques, of which colorimetry is commonly employed [20, 22]. The second path implies the presence of short-lived radical and atomic species: O, •NO, •OH, O2•−, as well as non-radical chemical compounds such as, e.g., singlet oxygen 1O2. Aside from creating direct oxidative stress , these species can regulate various cellular processes by, e.g., altering cellular uptake of metal ions . These short-lived radicals were also shown to initiate radical reactions in liquid media .
This chapter disserts methods of detection, identification and quantification of short-lived chemical species in solutions in contact with CAPs.
2. Detection of plasma-generated RONS in liquids
2.1 Hydroxyl and superoxide radicals
Upon interaction with water and oxygen moieties, CAPs generate hydroxyl radicals •OH and superoxide radical anions O2•−. These short-lived species possess highly oxidising and cytotoxic properties, and are suggested to be one of the main causes of biomedical activity of cold plasma [14, 28].
In aqueous systems, these radicals are often detected using optical methods. These methods usually employ induction of colour or degradation of dyes (colorimetry) . In this method, a coloured dye is degraded by plasma-generated species, and the loss of colour is quantified using UV-vis spectrophotometry. Some of the most commonly used dyes used to assess ROS are methylene blue and methylene red [29, 30], often used in research associated with CAPs for water treatment and pollutant removal [30, 31]. However, the main difficulty in employing this method is the non-specific degradation of such dyes. For example, decolouration of methylene blue not only occurs via reaction with •OH radicals but also with other CAP-induced species, e.g., ozone . Superoxide radical anions can also be detected by degradation of dyes  with the same limitations.
Similarly, induction/decay of fluorescent properties of chemical molecules is also used. Degradation of some fluorophores was used to detect oxygen-centred radicals on surfaces in contact with processing plasmas .
Terephthalic acid (TA) has been reported to detect •OH radicals in plasma-liquid systems [35, 36, 37]. The method is based on the induced fluorescence due to the formation of hydroxy-substituted TA (Scheme 1). This presents a simpler and more selective method of detection of the hydroxyl radicals, as demonstrated by Attri and co-workers . However, possible oxidation of terephthalic acid by RONS  is usually ignored. Another method based on aromatic hydroxylation was recently suggested by Zhang et al. Using salicylic acid as a substrate, the authors obtained mono- and disubstituted products, which were analysed by HPLC .
Other alternatives include different chemical probes, such as, e.g., dimethylsulphoxide, followed by quantification of the formed formaldehyde HCHO or methanesulfinic acid [35, 40] (Scheme 2). However, other reactions leading to both production and degradation of HCHO can occur in a plasma-liquid system, as was shown by Ma et al. . The kinetics of these complex chemical processes needs to be considered for quantitative assessment of the hydroxyl radical in liquids.
Shirai et al. observed chemiluminescence at the plasma-liquid interface when alkaline solutions of luminol were exposed to CAP, presumably due to reactions with •OH and O2•− . Bekeschus et al. compared the amount of superoxide produced by different plasmas in liquid using colorimetric analysis with cytochrome C .
Furthermore, the information on the availability of radical species in liquids exposed to CAPs is often obtained using electron paramagnetic resonance (EPR) spectroscopy. EPR is the most direct method of radical detection in liquids [43, 44]. The method is based on the detection of paramagnetic species, e.g., free radicals with an unpaired electron. In EPR, several methods of radical detection are used . Free radicals such as hydroxyl and superoxide are very short-lived and cannot be detected directly. Hence, spin probes and spin traps are employed. An example of a spin probe is 1-hydroxy-3-methoxycarbonyl-2,2,5,5-tetramethylpyrrolidine (CMH). CMH is a cyclic hydroxylamine, which reacts with superoxide radicals to form a nitroxide radical detected by EPR: =N▬OH → =N▬O• . However, this method is mostly used in biological systems rather than plasma-liquid systems. The selectivity is not explicitly known, and possible difficulties may arise from interferences by other plasma-induced ROS.
The second method is the formation of radical adducts in reactions of spin traps (organic molecules, usually nitrones) with free radicals (Scheme 3). The formed spin adducts are organic nitroxides with longer half-lives compared to the analysed radicals, and thus detectable by EPR [47, 48]. In the past decade, the spin trapping of CAP-induced radicals in liquids has gained vast attention. Numerous groups have performed detection of •OH and O2•−/•OOH radicals in aqueous media by EPR. Tani et al., Takamatsu et al. and Uchiyama et al. performed the detection of radical species in plasma-treated liquids by using various spin traps [49, 50, 51]. Tresp et al. assessed the concentrations of the •OH and O2•− radicals by monitoring the radicals adducts of 5-tert-butoxycarbonyl-5-methyl-1-pyrroline
The advantage of spin trapping over other methods is that the non-selectivity of a spin trap is not a drawback. Indeed, the characteristic features of the EPR signals are different for different radical adducts due to the hyperfine coupling: interaction with magnetic moments of nearby nuclei with non-zero spin numbers (14N, 1H, 13C, 17O, etc.). The hyperfine values of an adduct therefore depend on the structure of a specific adduct: DMPO-OH and DEPMPO-OH have different chemical structures and thus different features of their EPR spectra (see, e.g., [15, 44]). Also, if a spin trap (e.g., DEPMPO) forms adducts with several radicals such as •OH and O2•−, the amounts of both adducts can be obtained from the same EPR spectrum . This feature enables studying the source of the radicals produced by CAPs by using isotopically labelled water (H217O, 2H2O) to distinguish between the radicals formed from the gas phase water and the liquid water [15, 44, 54]. More recently, the use of PBN and DMPO spin traps helped identify the nature of the radicals generated by CAPs from organic solvents (chloroform and
However, despite being a very versatile and the most direct method of radical detection in liquids, spin trapping and EPR analysis have limitations, which need to be considered in experimental settings. Before proceeding to the detection of other short-lived species, we first address in the following section the factors limiting the applicability of spin trapping and EPR analysis in CAP-liquid systems.
2.2 Limitations of spin trapping
Quantification of the radical adduct (i.e., concentration in the analysed solution) can be achieved by calibrating an EPR spectrometer with solutions of a stable radical compound, e.g., a stable nitroxide (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) or its derivatives [15, 22, 44]. However, obtaining concentration values of radicals (rather than the formed adduct) in liquids is hindered due to various side reactions of the analysed radicals: recombination (e.g., •OH into H2O2), reactions with scavengers (solvated electrons, hydrogen atoms, possible physiological media components [22, 37, 42, 48]). Comparison of reaction rate coefficients of these processes needs to be performed. Otherwise, EPR with spin trapping provides only semi-quantitative data: the amounts of a radical adduct formed under different conditions follow the same trend as the initial concentration of the free radical.
Different spin traps have different affinity towards different radicals expressed by the rate coefficients of the respective reactions [44, 45]. Moreover, the stability of the formed radical adducts varies in orders of magnitude. Under physiological conditions (pH, temperature, etc.), the half-life of the radical adduct with superoxide radical of the DMPO spin trap is 45 s, with DEPMPO, it is 14 min, and with 3,5-dimethyl-5-(iso-propoxycarbonyl)-1-pyrroline
Hence, the choice of a spin trap is a very important factor, which may affect both the quantitative and the qualitative results. Selectivity, adduct stability and last but not least commercial availability of a spin trap should be taken into account when preparing for the analysis of radical species in liquids.
The limitations discussed above are known in biological systems, with limited production and diversity of RONS. With CAPs, various atomic and radical reactive species can be simultaneously delivered to the liquid. Our previous work showed that nitroxides can decay via reactions with the same species from which they were formed . Our results demonstrated that the loss of nitroxide moiety in plasma-exposed water occurred via reactions with •OH radicals, H atoms and oxygen species (atomic oxygen and/or ozone) with pseudo-first-order kinetics (Scheme 5). Hence, to perform even relative measurements of RONS induced in water by CAPs, a study of the product (spin adduct, aromatic substitution product, etc.) concentration development over time may be necessary to exclude possibilities of the increasing analyte degradation.
Another factor affecting the analysis is the potential decay of the nitrone spin traps themselves. Upon plasma exposure, it was shown that DEPMPO spin trap was partially degraded, yielding carbon-centred radicals, which were trapped by the remaining DEPMPO . Similarly, PBN can undergo degradation into
A number of ‘false-positive’ results have been identified for spin trapping. For instance, a nucleophilic addition via the Forrester-Hepburn mechanism, either direct  or metal-catalysed  (Scheme 6), may lead to the formation of nitroxides with the same structure as the radical adducts. It is thus important to perform control experiments with no CAP-induced RONS, to assess the possible interference from such reactions.
2.3 Solvated electrons
Many types of plasma set-ups either have discharges to the surface of the liquid (e.g., floating electrode plasmas with no gas flow), or an electron-rich afterglow (most plasma jets, with the exception of COST jet-type set-ups) [15, 16, 23]. In such cases, not just the RONS, but the electrons too may interact with the CAP-exposed liquid [16, 62]. Solvated electrons contribute to the additional charge in the liquid and induce electrochemical reactions , affecting potential substrates.
Rumbach et al. have demonstrated that both electron transfer and neutral reactions occur when CAPs interact with aqueous media. The authors report an optical technique employing a series of individual diode lasers for the spectroscopic detection of solvated electrons [63, 64]. Another (chemical) approach describes the analysis of the pH of the saline solutions in an electrochemical system where a cathode is substituted by a plasma jet. In such system, the pH changes as a result of the chlor-alkali process initiated by plasma electrons . To the best of our knowledge, no other works describe chemical detection of solvated electrons in plasma-liquid systems.
Several reviews and reports describe possibilities of chemical detection of solvated electrons [66, 67]. The methods include indirect detection, e.g., EPR analysis of the DMPO-spin trapped benzyl radicals, which were formed upon reaction of benzyl chloride with solvated electrons . These methods could prove very useful in plasma-liquid systems, although their direct applicability (limitations due to selectivity, etc.) needs to be determined.
2.4 Atomic oxygen, singlet oxygen and ozone
Some of the most reactive and biologically relevant species created by CAPs are atomic oxygen O, ozone O3 and singlet oxygen 1O2 [14, 69]. It has been shown by Benedikt and co-workers that oxygen atoms are delivered to exposed liquid solutions from the gas phase plasma, where they are generated [70, 71], while computational results indicate that these extremely reactive atomic species can also be formed inside the liquid from hydroxyl radicals . Singlet oxygen and ozone are generally considered short-lived (compared to e.g. hydrogen peroxide), although they are more stable than the atomic or radical species.
Ozone is often detected colorimetrically using the degradation of coloured dyes, e.g., methylene blue. This method of ozone detection is highly non-selective, since the dyes can be degraded by other CAP-produced RONS, including •OH radicals . Kovačević et al. measured ozone delivery to the plasma-exposed liquid using iodometry, while the solubilised ozone remaining in the liquid after the CAP treatment was detected with decolourisation of indigo trisulphonate . However, recently, Tarabová et al. reported that indigo dyes decay in reactions with other RONS, e.g., secondary •OH radicals produced in liquids after the plasma exposure . The non-selectivity of the iodometric method is due to the other oxidising RONS . Fluorescent probes have also been used to detect ozone in plasma-liquid systems, albeit not without selectivity issues .
Benedikt and co-workers detected oxygen atoms in aqueous solutions of phenol, with further MS analysis of the formed hydroxylated products [70, 71]. However, hydroxylation of phenol can also occur in a reaction with hydroxyl radicals [73, 75].
EPR detection of a combination of CAP-induced O, 1O2 and O3 in aqueous media was performed by Takamatsu et al. The authors used 2,2,5,5-tetramethyl-3-pyrroline-3-carboxamide (TPC) as a chemical detector of oxygenated species in liquid . TPC is oxidised to produce a stable nitroxide, which can be detected by EPR. While the oxidation of TPC was non-selective, the addition of sodium azide NaN3 as a singlet oxygen scavenger allowed to distinguish between the TPC oxidised by 1O2 and by all other species.
We have previously used 2,2,6,6-tetramethylpiperidine (TEMP), another cyclic amine, to detect O/1O2/O3 produced by an atmospheric pressure plasma via its oxidation to 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO) [22, 44]. Adding NaN3 let us individually assess the concentrations of 1O2 and O/O3 induced by CAPs (Scheme 7). Our work showed that although TEMPO was not produced in reactions with H2O2 or O2•−, it could be formed by other plasma-induced RONS: ozone and possibly by atomic oxygen .
Later, Elg et al. demonstrated that most TEMPO was in fact formed by atomic oxygen by comparing ozone densities in the gas phase with concentrations of the formed TEMPO . However, the main contributor to the production of a stable nitroxide in these reactions would depend on the densities of O and O3 (and therefore, a specific plasma set-up) in each case. Other limitations of this method are related to the loss reactions of nitroxides , as described above (see Scheme 5).
2.5 Nitrogen oxides and peroxynitrite
Nitric oxide •NO is considered one of the CAP-induced RONS responsible for bactericidal effects and wound healing. Furthermore, its reactions with and within aqueous media produce a large variety of secondary RONS . Unlike the more persistent products of the •NO transformations such as NO2− and NO3−, which are detected colorimetrically [20, 48, 72, 73], the radical nitric oxide itself is monitored using spin trapping and EPR. Many EPR methods of detection of nitric oxide are known in biological milieu. Among these are the use of dithiocarbamate metal complexes, oxidation of nitronyl nitroxides, etc. .
In plasma-liquid systems, •NO has been detected intracellularly using fluorescent probes [51, 79] and directly in media by EPR with iron complexes of
Another method to detect nitric oxide in CAP-liquid systems is the transformation of nitronyl nitroxides such as 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide (PTIO) or its derivatives [48, 51, 58]. PTIO, a stable nitroxide radical, reacts with nitric oxide to form 2-phenyl-4,4,5,5-tetramethylimidazoline 1-oxyl (PTI), an imino nitroxide radical (Scheme 9). EPR analysis with deconvolution of the spectra via radical signal simulations allows differentiating between the two radicals [48, 78]. We previously showed that the limitations of the method are related to the nitroxide decay pathways (see above). The other issue is the reverse transformation (oxidation) of PTI to PTIO in oxygen-containing plasmas, even in the absence of nitrogen . This makes detection of nitric oxide in plasma-liquid systems an extremely difficult task, when both the absence of the detectable •NO (PTIO) and its presence (a ‘false positive’ with dithiocarbamates) can be due to the limitations of each method.
Lukes et al. used nitrosation and nitration reactions of phenol as a technique to detect post-plasma discharge formation of nitrogen oxides (•NO and •NO2) in aqueous media  (Scheme 10). It can be used to detect these two radicals directly during plasma exposure, or the secondary radicals formed during the decay of peroxynitrite (see below). As far as we know, no other methods of •NO2 detection have been reported specifically in plasma-treated solution, although other ways of detecting nitrogen dioxide in aqueous media are available in literature: e.g., spin trapping with nitrones  or nitroalkanes .
The peroxynitrite anion ONOO− is another reactive species induced by CAPs in water-based media [14, 72, 83]. Lukes et al. were the first to evaluate its formation and stability in aqueous solutions exposed to plasma. Peroxynitrite decayed rapidly in acidic conditions via peroxynitrous acid decomposition into •OH and •NO2, which were detected via phenol derivatisation . The instability of peroxynitrite in neutral pH was later demonstrated by Weltmann and co-workers [84, 85]. Here, peroxynitrite was detected by exposing solutions of L-tyrosine to CAP and further MS analysis of the formed 3-nitrotyrosine. The authors have also assessed the interferences from the •NO radicals by introducing an •NO donor . When L-tyrosine was added to plasma-treated solutions after the exposure, no nitration product was detected, suggesting the short-lived nature of peroxynitrite under the applied conditions. Girard et al. developed a method of direct spectrophotometrical detection of peroxynitrite based on its absorption properties in the UV region . Here, peroxynitrite was only detected in highly basic solutions, confirming its short life at neutral pH. Xu et al. used 1-hydroxy-2,2,6,6-tetramethyl-4-oxo-piperidine (a hydroxylamine) to detect both superoxide radical anion and peroxynitrite anion via EPR analysis of the formed 4-oxo-2,2,6,6-tetramethylpiperidine 1-oxyl  (distinguishing between the two is not possible with this method [46, 87]).
2.6 Hypochlorite formed in physiological media
Both the physiological media used in research in vitro (PBS, DMEM, etc.) and the blood and tissues in vivo have high concentration of chloride salts. Chloride anions can act as scavengers of the CAPs-generated RONS in media, e.g., the •OH radicals [28, 88]. It has been suggested by different research groups that chlorinated species such as hypochlorite ClO− (short-lived under physiological conditions ) can be formed in solutions upon interaction with the plasma RONS [79, 85, 89]. Wende et al. have assessed the effects of the potential ClO− presence in situ on the overall biomedical efficacy of CAPs . Recently, Piskarev et al. showed that the ClO− anion is formed in plasma-treated saline solutions via reactions of the Cl− anion with superoxide or hydroxyl radicals from plasma .
The detection of hypochlorite in solutions was performed by direct UV detection  or by exposure of a solution of L-tyrosine to CAPs [84, 85]. The latter method is based on the MS detection of the chlorinated product of L-tyrosine. The UV detection requires large concentrations of hypochlorite, while the MS method is highly sensitive with a low detection limit. However, Bekeshus et al. detected no hypochlorite formation with their plasma set-up . The formation of hypochlorite (formed from the initial CAP-generated RONS in liquid media), therefore, depends on the particular plasma set-up and application conditions.
3. Conclusion and perspectives
A plethora of methods are available for the detection of reactive oxygen and nitrogen species induced in liquids by plasma. Chemical ‘detector’ systems based on colorimetry, fluorescence, (LC-)MS analysis of the products and EPR spin trapping are some of the techniques used in the detection of atomic, radical, molecular and ionic short-lived RONS. Each method has its potential and limitations; the latter associated with the decay of the products, low selectivity and other factors.
Despite several limitations, spin trapping coupled with EPR analysis as an analytical method of radical detection in CAPs systems has a very high value as the most direct method of radical detection in liquids. Aside from the liquid media itself, it has a potential to be used in plasma-gel systems, which mimic interaction of CAPs with tissue , if gels are formed from nitrone molecules . The availability of EPR equipment is not necessarily crucial to perform spin trapping of RONS. An interesting alternative to EPR analysis of radical adducts was demonstrated by Guo et al. and Tuccio et al., who used liquid chromatography and mass spectrometry systems to analyse the adducts of oxygen-centred radicals of the DMPO and DEPMPO spin traps [93, 94].
The emerging role of the chlorinated species in CAP-treated media is gaining attention. ClO− anion, a biomedically relevant species, has been monitored using UV absorption spectroscopy and modification of tyrosine. However, other detection methods may need to be used, especially in cases when hypochlorite is not detected. For example, colorimetric analysis on the oxidation of 3,3′,5,5′-tetramethylbenzidine is a simple technique , albeit with yet unknown limitations due to the presence of other oxidising RONS. Moreover, the highly oxidative nature of the hypochlorite anion in plasma-treated physiological solutions is often emphasised, but the possibility of formation of other anions such as ClO2−, ClO3− and ClO4− has not been addressed. Since these species are cytotoxic, monitoring them in physiological media exposed to plasma can provide valuable information on the biomedical effects of CAPs. Possible analyses can include ion chromatography [96, 97].
Finally, some other short-lived yet highly reactive species are overlooked in the current research. Among these is the carbonate radical anion CO3•.: a very potent oxidising agent causing DNA damage , which can be formed in a reaction of peroxynitrite with ambient CO2 [99, 100]. The determination of this and other reactive species will aid in completing the picture of the plasma-produced ‘cocktail’ of reactive species. It can facilitate both the understanding of the existing CAP devices and their effects, and the development of new plasma systems with dedicated RONS concentrations for specific applications.
Yury Gorbanev thanks the European Marie Sklodowska-Curie Individual Fellowship within Horizon2020 (‘LTPAM’, grant no. 743151) for supporting his work.
Conflict of interest
The authors have no conflict of interest to declare.