Spectral Probe for Electron Transfer and Addition Reactions of Azide Radicals with Substituted Quinoxalin-2-Ones in Aqueous Solutions

The azide radical (N3●) is one of the most important one-electron oxidants used extensively in radiation chemistry studies involving molecules of biological significance. Generally, it was assumed that N3● reacts in aqueous solutions only by electron transfer. However, there were several reports indicating the possibility of N3● addition in aqueous solutions to organic compounds containing double bonds. The main purpose of this study was to find an experimental approach that allows a clear assignment of the nature of obtained products either to its one-electron oxidation or its addition products. Radiolysis of water provides a convenient source of one-electron oxidizing radicals characterized by a very broad range of reduction potentials. Two inorganic radicals (SO4●−, CO3●−) and Tl2+ ions with the reduction potentials higher, and one radical (SCN)2●− with the reduction potential slightly lower than the reduction potential of N3● were selected as dominant electron-acceptors. Transient absorption spectra formed in their reactions with a series of quinoxalin-2-one derivatives were confronted with absorption spectra formed from reactions of N3● with the same series of compounds. Cases, in which the absorption spectra formed in reactions involving N3● differ from the absorption spectra formed in the reactions involving other one-electron oxidants, strongly indicate that N3● is involved in the other reaction channel such as addition to double bonds. Moreover, it was shown that high-rate constants of reactions of N3● with quinoxalin-2-ones do not ultimately prove that they are electron transfer reactions. The optimized structures of the radical cations (7-R-3-MeQ)●+, radicals (7-R-3-MeQ)● and N3● adducts at the C2 carbon atom in pyrazine moiety and their absorption spectra are reasonably well reproduced by density functional theory quantum mechanics calculations employing the ωB97XD functional combined with the Dunning’s aug-cc-pVTZ correlation-consistent polarized basis sets augmented with diffuse functions.


Introduction
The azide radical (N 3 • ) is one of the most important one-electron oxidants used extensively in radiation chemistry studies involving inorganic [1][2][3], and aromatic compounds [4][5][6][7][8][9], and also molecules of biological significance [10][11][12][13][14][15]. In general, it is assumed that N 3 • reacts in aqueous solution only by electron transfer [4,16,17]. Therefore, the value of the standard reduction potential of the N 3 • /N 3 − redox couple (E 0 = 1.33 ± 0.01 V vs. NHE) [18], which applies to reactions in aqueous solutions, is essential for understanding the mechanism and the kinetics of oxidation of organic compounds by N 3 • . Its oxidation reactions are particularly rapid, even more rapid than the reactions involving some stronger oxidants such as Br 2 •− and CO 3 •− [19]. This is probably due to the high self-exchange rate constant of ≈4 × 10 4 M −1 s −1 for the N 3 • /N 3 − couple inferred from the cross-relationship of Marcus theory [16].
These observations are in line with the lower reduction potential of 2-TU (≈ +0.7 V vs. NHE) in comparison to N 3 • and therefore in such designed systems, oxidation of 2-TU leads directly to 2TU + (S • ) radical cations.
Interestingly, there are several reports indicating the possibility of N 3 • addition in aqueous solutions to organic compounds containing double bonds. Spin-trapping experiments performed for the N 3 • detection in aqueous solutions of an azide/catalase/H 2 O 2 and an azide/peroxidase/H 2 O 2 using phenyl-tert-butyl nitrone (PBN) and 5,5-dimetyl-1-pyrroline-N-oxide (DMPO) confirmed the presence PBN/DMPO-N 3 radical adducts which were detected by ESR techniques. They are formed via reactions represented by (6) and (7), respectively. By using 14 N-and 15 N-labelled NaN3 it was possible to confirm unequivocally that N3 • added to the C=N bond in PBN [33].
PBN-N3 radical adducts were detected in similar spin-trapping experiments performed on photolyzed aqueous solutions containing azide/H2O2 [34], and azido cobalt(III) complexes [35]. Similarly, DMPO-N3 radical adducts were also detected in aqueous solutions of azide/cytochrome c oxidase and azide/cytochrome c oxidase/H2O2 systems [36], azide/H2O2-activated endogeneous cytochrome c peroxidase [37], and in a photolyzed azide/H2O2 system [34]. These experiments clearly showed the possibility for another reaction channel of N3 • in aqueous solutions such as addition to double bonds (Equations (6) and (7)). However, they do not provide us with information related to the reactivity of N3 • by processes other than electron transfer. Such knowledge is important for evaluating the usefulness of N3 • as a secondary oxidant in biological studies.
It should be noted that the standard reduction potential of the N3 • /N3 − redox couple decreases in going from polar to less polar solvents [38]. Therefore N3 • is not as strong oxidizing agent in such solvents as it is in water. This leaves open the possibility for other reaction channels such as addition to double bonds or hydrogen abstraction. For instance, the absolute rate constants for the addition reactions of N3 • with a series of ring-substituted styrenes (p-CF3, m-CF3, p-Cl, H, m-CH3, p-CH3, p-CH3O) in acetonitrile were found to vary between 1 × 10 6 M −1 s −1 and 5 × 10 7 M −1 s −1 . A correlation of log(kadd) with Hammett σ + constants yields  + = −1.2 which indicates the electrophilic nature of N3 • [39]. On the other hand, the rate constants for the reaction of N3 • with α-and β-substituted styrenes and simple alkyl-and alkoxy-substituted olefins which vary between 1 × 10 6 M −1 s −1 and 1 × 10 9 M −1 s −1 were nicely correlated with the corresponding ionization potentials (IPe). The negative slope of the linear plot of log(kadd) versus IPe indicates that these reactions are occurring with considerable charge transfer interactions in the transition state and are dominated by polar effects [39].
With these premises, in the current studies we report our investigations on the reaction of N3 • with the quinoxalin-2-one derivatives which are multifunctional molecules consisting of an aromatic ring connected in neighboring positions with a heterocyclic ring, in aqueous solutions. We selected six 3-methyl-1H-quinoxalin-2-one derivatives, five of them substituted in position 7 with various electron donating (-OCH3, -NH2) and electron withdrawing (-F, -CN, -CF3) groups ( Figure 1).   (6) and (7), respectively. By using 14 N-and 15 N-labelled NaN3 it was possible to confirm unequivocally that N3 • added to the C=N bond in PBN [33].
PBN-N3 radical adducts were detected in similar spin-trapping experiments performed on photolyzed aqueous solutions containing azide/H2O2 [34], and azido cobalt(III) complexes [35]. Similarly, DMPO-N3 radical adducts were also detected in aqueous solutions of azide/cytochrome c oxidase and azide/cytochrome c oxidase/H2O2 systems [36], azide/H2O2-activated endogeneous cytochrome c peroxidase [37], and in a photolyzed azide/H2O2 system [34]. These experiments clearly showed the possibility for another reaction channel of N3 • in aqueous solutions such as addition to double bonds (Equations (6) and (7)). However, they do not provide us with information related to the reactivity of N3 • by processes other than electron transfer. Such knowledge is important for evaluating the usefulness of N3 • as a secondary oxidant in biological studies.
It should be noted that the standard reduction potential of the N3 • /N3 − redox couple decreases in going from polar to less polar solvents [38]. Therefore N3 • is not as strong oxidizing agent in such solvents as it is in water. This leaves open the possibility for other reaction channels such as addition to double bonds or hydrogen abstraction. For instance, the absolute rate constants for the addition reactions of N3 • with a series of ring-substituted styrenes (p-CF3, m-CF3, p-Cl, H, m-CH3, p-CH3, p-CH3O) in acetonitrile were found to vary between 1 × 10 6 M −1 s −1 and 5 × 10 7 M −1 s −1 . A correlation of log(kadd) with Hammett σ + constants yields  + = −1.2 which indicates the electrophilic nature of N3 • [39]. On the other hand, the rate constants for the reaction of N3 • with α-and β-substituted styrenes and simple alkyl-and alkoxy-substituted olefins which vary between 1 × 10 6 M −1 s −1 and 1 × 10 9 M −1 s −1 were nicely correlated with the corresponding ionization potentials (IPe). The negative slope of the linear plot of log(kadd) versus IPe indicates that these reactions are occurring with considerable charge transfer interactions in the transition state and are dominated by polar effects [39].
With these premises, in the current studies we report our investigations on the reaction of N3 • with the quinoxalin-2-one derivatives which are multifunctional molecules consisting of an aromatic ring connected in neighboring positions with a heterocyclic ring, in aqueous solutions. We selected six 3-methyl-1H-quinoxalin-2-one derivatives, five of them substituted in position 7 with various electron donating (-OCH3, -NH2) and electron withdrawing (-F, -CN, -CF3) groups ( Figure 1).  PBN-N 3 radical adducts were detected in similar spin-trapping experiments performed on photolyzed aqueous solutions containing azide/H 2 O 2 [34], and azido cobalt(III) complexes [35]. Similarly, DMPO-N 3 radical adducts were also detected in aqueous solutions of azide/cytochrome c oxidase and azide/cytochrome c oxidase/H 2 O 2 systems [36], azide/H 2 O 2 -activated endogeneous cytochrome c peroxidase [37], and in a photolyzed azide/H 2 O 2 system [34]. These experiments clearly showed the possibility for another reaction channel of N 3 • in aqueous solutions such as addition to double bonds (Equations (6) and (7)). However, they do not provide us with information related to the reactivity of N 3 • by processes other than electron transfer. Such knowledge is important for evaluating the usefulness of N 3 • as a secondary oxidant in biological studies. It should be noted that the standard reduction potential of the N 3 • /N 3 − redox couple decreases in going from polar to less polar solvents [38]. Therefore N 3 • is not as strong oxidizing agent in such solvents as it is in water. This leaves open the possibility for other reaction channels such as addition to double bonds or hydrogen abstraction. For instance, the absolute rate constants for the addition reactions of N 3 • with a series of ring-substituted styrenes (p-CF 3 , m-CF 3 , p-Cl, H, m-CH 3 , p-CH 3 , p-CH 3 O) in acetonitrile were found to vary between 1 × 10 6 M −1 s −1 and 5 × 10 7 M −1 s −1 . A correlation of log(k add ) with Hammett σ + constants yields ρ + = −1.2 which indicates the electrophilic nature of N 3 • [39]. On the other hand, the rate constants for the reaction of N 3 • with αand β-substituted styrenes and simple alkyl-and alkoxy-substituted olefins which vary between 1 × 10 6 M −1 s −1 and 1 × 10 9 M −1 s −1 were nicely correlated with the corresponding ionization potentials (IP e ). The negative slope of the linear plot of log(k add ) versus IP e indicates that these reactions are occurring with considerable charge transfer interactions in the transition state and are dominated by polar effects [39].
With these premises, in the current studies we report our investigations on the reaction of N 3 • with the quinoxalin-2-one derivatives which are multifunctional molecules consisting of an aromatic ring connected in neighboring positions with a heterocyclic ring, in aqueous solutions. We selected six 3-methyl-1H-quinoxalin-2-one derivatives, five of them substituted in position 7 with various electron donating (-OCH 3 , -NH 2 ) and electron withdrawing (-F, -CN, -CF 3 ) groups ( Figure 1).
x FOR PEER REVIEW 3 of 23 (6) and (7), respectively. By using 14 N-and 15 N-labelled NaN3 it was possible to confirm unequivocally that N3 • added to the C=N bond in PBN [33].
PBN-N3 radical adducts were detected in similar spin-trapping experiments performed on photolyzed aqueous solutions containing azide/H2O2 [34], and azido cobalt(III) complexes [35]. Similarly, DMPO-N3 radical adducts were also detected in aqueous solutions of azide/cytochrome c oxidase and azide/cytochrome c oxidase/H2O2 systems [36], azide/H2O2-activated endogeneous cytochrome c peroxidase [37], and in a photolyzed azide/H2O2 system [34]. These experiments clearly showed the possibility for another reaction channel of N3 • in aqueous solutions such as addition to double bonds (Equations (6) and (7)). However, they do not provide us with information related to the reactivity of N3 • by processes other than electron transfer. Such knowledge is important for evaluating the usefulness of N3 • as a secondary oxidant in biological studies.
It should be noted that the standard reduction potential of the N3 • /N3 − redox couple decreases in going from polar to less polar solvents [38]. Therefore N3 • is not as strong oxidizing agent in such solvents as it is in water. This leaves open the possibility for other reaction channels such as addition to double bonds or hydrogen abstraction. For instance, the absolute rate constants for the addition reactions of N3 • with a series of ring-substituted styrenes (p-CF3, m-CF3, p-Cl, H, m-CH3, p-CH3, p-CH3O) in acetonitrile were found to vary between 1 × 10 6 M −1 s −1 and 5 × 10 7 M −1 s −1 . A correlation of log(kadd) with Hammett σ + constants yields  + = −1.2 which indicates the electrophilic nature of N3 • [39]. On the other hand, the rate constants for the reaction of N3 • with α-and β-substituted styrenes and simple alkyl-and alkoxy-substituted olefins which vary between 1 × 10 6 M −1 s −1 and 1 × 10 9 M −1 s −1 were nicely correlated with the corresponding ionization potentials (IPe). The negative slope of the linear plot of log(kadd) versus IPe indicates that these reactions are occurring with considerable charge transfer interactions in the transition state and are dominated by polar effects [39].
With these premises, in the current studies we report our investigations on the reaction of N3 • with the quinoxalin-2-one derivatives which are multifunctional molecules consisting of an aromatic ring connected in neighboring positions with a heterocyclic ring, in aqueous solutions. We selected six 3-methyl-1H-quinoxalin-2-one derivatives, five of them substituted in position 7 with various electron donating (-OCH3, -NH2) and electron withdrawing (-F, -CN, -CF3) groups ( Figure 1).  Quinoxalin-2-one derivatives are also ubiquitous in biology and have recently received much attention connected with their biological properties and pharmaceutical applications [40,41]. A key factor that is decisive in their biological activity is substitution at the carbon-3 in the pyrazine ring and at carbons 6 and/or 7 in the benzene ring. Nearly all biologically active derivatives are substituted in those specific positions. The application of quinoxaline derivatives is strongly related to possible one-electron redox processes involving the quinoxaline moiety. There are several comprehensive reviews covering quinoxaline chemistry and applications [40,[42][43][44][45][46][47]. Nonetheless, studies devoted to quinoxaline derived radicals are rather scarce and include studies performed by pulse radiolysis [48][49][50][51][52], photochemistry [53][54][55][56][57], electrochemistry [58][59][60], and Fenton systems [61].
In principle, it is well known that reduction potentials are linearly dependent on Brown's σ p + values as it was shown for benzene radical cations [62], and 4-substituted aniline radical cations [63]. In other words, electron donating substituents lower the oneelectron reduction potential of a given redox couple. Taking into account Brown's σ p + values [64], one can expect the highest reduction potential for the 7-CN-3MeQ derivative and the lowest reduction potential for the 7-NH 2 -3-MeQ derivative. This conclusion is also in line with reduction potential predictions, based on accurate quantum chemical methods (DFT), of quinoxaline and a number of its derivatives with electron-donating and electron-withdrawing substituent groups [65].
Structure Activity Relationship (SAR) studies revealed that quinoxalin-2-ones derivatives are bound in very specific positions in proteins [66,67]. Therefore, they or radicals derived from them can interact with either amino acid residues or radicals derived from them. For example, certain amino acid residues-TyrOH, TrpNH, and CysSH-are particularly vulnerable to oxidation. Therefore, the radical cations derived from quinoxalin-2-one derivatives can potentially oxidize them to TyrO • , TrpN • and CysS • radicals, respectively. On the other hand, these radicals are reasonably good electron acceptors and can act as oxidants of quinoxalin-2-one derivatives intercalated in a protein matrix.
These examples strongly indicate a need to obtain a comprehensive and systematized, and at the same time reliable knowledge about spectral and kinetic properties of radicals and radical ions derived from quinoxalin-2-ones. Since N 3 • is both commonly used as an one-electron oxidant of biologically important compounds, and yet has the confirmed possibility of adding to double bonds, there is a strong need to unambiguously assign observed transient absorption spectra to either its one-electron oxidation products or its addition products. Therefore, the main aim of this current work was to find an experimental approach that would allow us to clearly define the nature of the obtained products.
Radiolysis of water provides a very convenient source of one-electron oxidizing radicals characterized by a very broad range of reduction potentials which is very useful for studying oxidation reactions of molecules of biological significance [68,69]. In the current study, we selected two inorganic radicals (SO 4 •− , CO 3 •− ) and Tl 2+ ions, with reduction potentials higher than the reduction potential of N 3 • , and one radical (SCN) 2 •− with a nearly equal, however lower, reduction potential as that of N 3 • (vide Table S1). These radicals and Tl 2+ ions act predominantly as electron acceptors, however, the radicals can also react by hydrogen abstraction or addition reactions [70][71][72][73][74][75][76][77]. These latter reactions are generally very slow and might not be observed by pulse radiolysis [19].
The basic idea of this experimental study is that we have probed absorption spectra formed in reactions of one-electron oxidants with a series of quinoxalin-2-one derivatives and have confronted these spectra with absorption spectra formed from reactions of azide radicals with the same series of compounds. The mere fact that reactions of N 3 • with quinoxalin-2-ones take place with high rate constants does not ultimately prove that they are electron transfer reactions [50]. Cases, in which the absorption spectra formed in reactions involving N 3 • differ from the absorption spectra formed in the reactions involving other one-electron oxidants, may indicate that N 3 • does not act as electron acceptor.

Absorption Spectra Recorded after Reaction of N 3
• and SO 4 •− with 3-MeQ At the beginning, we selected for comparative studies the sulfate radical anion (SO 4 •− ) with the highest reduction potential (E 0 = +2.44 V vs. NHE) which should be high enough to oxidize 3-MeQ, and is also much higher that the reduction potential of N 3 • (E 0 = +1.33 V vs. NHE). Provided that the reduction potential of N 3 • (E 0 = +1.33 V vs. NHE) is high enough to oxidize 3-MeQ, one should expect similar oxidation products formed from the reaction of these two radicals with 3-MeQ.
Under conditions ensuring participation in the oxidation reaction of only SO 4 •− (vide Materials and Methods), the transient absorption spectrum recorded at 10 µs after the electron pulse in aqueous solutions at pH = 7 and containing 0.1 mM 3-MeQ is characterized by two absorption maxima at λ max = 375 and 480 nm and featureless absorption band which shows no distinct λ max > 270 nm ( Figure 2). This absorption spectrum was tentatively assigned by us to the product of one-electron oxidation of 3-MeQ by SO 4 •− . On the other hand, under conditions ensuring participation in the potential oxidation reaction of only N 3 • (vide Materials and Methods), the transient absorption spectrum recorded at 3 µs after the electron pulse in aqueous solutions at pH = 7 and containing 0.1 mM 3-MeQ is characterized by a narrow and distinct absorption band with λ max = 355 nm ( Figure 2). At the beginning, we selected for comparative studies the sulfate radical anion (SO4 •− ) with the highest reduction potential (E 0 = +2.44 V vs. NHE) which should be high enough to oxidize 3-MeQ, and is also much higher that the reduction potential of N3 • (E 0 = +1.33 V vs. NHE). Provided that the reduction potential of N3 • (E 0 = +1.33 V vs. NHE) is high enough to oxidize 3-MeQ, one should expect similar oxidation products formed from the reaction of these two radicals with 3-MeQ.
Under conditions ensuring participation in the oxidation reaction of only SO4 •− (vide Materials and Methods), the transient absorption spectrum recorded at 10 μs after the electron pulse in aqueous solutions at pH = 7 and containing 0.1 mM 3-MeQ is characterized by two absorption maxima at λmax = 375 and 480 nm and featureless absorption band which shows no distinct λmax > 270 nm ( Figure 2). This absorption spectrum was tentatively assigned by us to the product of one-electron oxidation  Due to depletion of the 3-MeQ ground state which absorbs at the spectral region < 380 nm (vide Figure S1) the observed absorption maxima in this spectral region were shifted towards shorter wavelengths in both systems studied (vide Figure S2). Despite this inconvenience, there is no doubt that reactions of 3MeQ with SO4 •− and N3 • lead to different primary transient products.
Interestingly, no reaction of (SCN)2 •− with 3-MeQ was observed in N2O-saturated aqueous solutions at pH 7 containing 1 mM 3-MeQ and 10 mM KSCN.  Due to depletion of the 3-MeQ ground state which absorbs at the spectral region <380 nm (vide Figure S1) the observed absorption maxima in this spectral region were shifted towards shorter wavelengths in both systems studied (vide Figure S2). Despite this inconvenience, there is no doubt that reactions of 3MeQ with SO 4 •− and N 3 • lead to different primary transient products.

Absorption Spectra Recorded after Reaction of N 3 • and SO 4 •− with 7-R-3-MeQ Derivatives
In order to check whether reactions of N 3 • and SO 4 •− lead again to different transient products, similar experiments were performed with 3-MeQ derivatives containing electronwithdrawing substituents (-CN, -CF 3 , and -F), and electron-donating substituent (-OCH 3 ) in position 7 in the benzene moiety.
The transient absorption spectra resulting from the reaction of SO 4 •− and recorded at 6-7 µs after the electron pulse in aqueous solutions at pH = 4 and containing 0.1 mM 7-R-3-MeQ derivatives with electron-withdrawing substituents are characterized by the similar features in comparison to the absorption spectrum recorded in the case of 3-MeQ.
They consisted of two absorption bands with maxima located at λ = 390 and 480 nm and featureless absorption band which shows no distinct λ max > 300 nm ( Figure 3a). Again, the transient absorption spectra resulting from the reaction of N 3 • and recorded at 6-8 µs after the electron pulse in aqueous solutions at pH = 7 and containing 0.1 mM 7-R-3-MeQ derivatives were characterized by similar features in comparison to the absorption spectrum recorded in the case of 3-MeQ. They were characterized by a narrow and distinct absorption band with λ max ≈ 350-360 nm (Figure 3b).    Similar spectral features were observed in the transient absorption spectrum resulting from the reaction of N 3 • with 7-NH 2 -3-MeQ recorded at 10 µs after the electron pulse, in aqueous solutions at pH = 7 and containing 0.1 mM 7-NH 2 -3-MeQ (vide Figure S3).

Absorption Spectra Recorded in Slightly Acidic and Neutral Solutions
In order to check whether the other one-electron oxidant (Tl 2+ ) is able to oxidize 7-OCH 3 -3-MeQ, the transient absorption spectra were recorded in N 2 O-saturated aqueous solutions containing 0.1 mM 7-OCH 3 -3-MeQ and 5 mM TlCl. The transient absorption spectrum resulted from the reaction of Tl 2+ with 7-OCH 3 -3-MeQ and recorded at 40 µs after the electron pulse, in aqueous solutions at pH 3.0 is characterized by the similar spectral features (Figure 4a) in comparison to the spectra resulted from the reaction of SO 4 •− and N 3 • . Therefore, these absorption spectra were assigned to the product of one-electron oxidation of 7-OCH 3 -3-MeQ which most likely is the 7-OCH 3 -3-MeQ •+ or its deprotonated neutral form 7-OCH 3 -3-MeQ • (vide Discussion).

Absorption Spectra Recorded in Alkaline Solutions
In the investigated pH range (3 to 11.3) 7-R-3-MeQ quinoxalin-2-ones exist in two forms ( Figure S4) that are involved in acid-base equilibria with the respective pK a values located between pH 8.3 and 10.0 (Table S2). In another words, in slightly acidic or neutral solutions the protonated form dominates whereas in alkaline solutions the anionic form.
Experiments, analogous to those in slightly acidic or neutral solutions, were performed in alkaline solutions in order to check whether the transient absorption spectra from reactions of anionic form of 7-OCH 3 -3-MeQ with N 3 • as compared with another oneelectron oxidant (CO 3 •− ) were characterized by the similar spectral features. The transient absorption spectra resulting from the reaction of N 3 • and CO 3 •− and recorded at 9 and 50 µs, respectively after the electron pulse in aqueous solutions at pH 11.3 and containing 0.1 mM 7-OCH 3 -3-MeQ are characterized by the similar features. They consisted of two absorption bands with maxima located at λ = 420 and 540 nm and the third absorption band with λ max = 320 nm (Figure 4b).

Reactions of N 3
• with 3-Methyl-1H-Quinoxalin-2-one (3-MeQ) and Its 7-Substituted Derivatives In order to determine directly the bimolecular rate constants of N 3 • with 3-MeQ and 7-R-MeQ, a kinetic analysis at various concentration of 3-MeQ and 7-R-MeQ (0.05 mM-0.5 mM) was performed. The growth kinetics were recorded at λ = 370 nm. The rate of formation, followed at that wavelength fits to a single exponential (Figure 5a,b).
The pseudo-first-order rate constants of the formation of the 370-nm absorption bands were plotted as a function of 3-MeQ and 7-CN-3-MeQ concentrations ( Figure 6). It is clearly seen that the pseudo-first-order rate constants measured at λ = 370 nm show a linear dependence on the concentration of 3-MeQ and 7-CN-3-MeQ in the full range of concentration studied. The slopes represent the second-order rate constants for the formation of transient resulting from the reaction of N 3 • with 3-MeQ and 7-CN-3-MeQ. Interestingly, the linear plots have non-zero intercepts that indicate the involvement of equilibria (vide Discussion) and that represent first-order rate constants for the backward reactions in the equilibria. The pseudo-first-order rate constants of the formation of the 370-nm absorption bands were plotted as a function of 3-MeQ and 7-CN-3-MeQ concentrations ( Figure 6). It is clearly seen that the pseudo-first-order rate constants measured at λ = 370 nm show a linear dependence on the concentration of 3-MeQ and 7-CN-3-MeQ in the full range of concentration studied. The slopes represent the second-order rate constants for the formation of transient resulting from the reaction of N3 • with 3-MeQ and 7-CN-3-MeQ. Interestingly, the linear plots have non-zero intercepts that indicate the involvement of equilibria (vide Discussion) and that represent first-order rate constants for the backward reactions in the equilibria.   The pseudo-first-order rate constants of the formation of the 370-nm absorption bands were plotted as a function of 3-MeQ and 7-CN-3-MeQ concentrations ( Figure 6). It is clearly seen that the pseudo-first-order rate constants measured at λ = 370 nm show a linear dependence on the concentration of 3-MeQ and 7-CN-3-MeQ in the full range of concentration studied. The slopes represent the second-order rate constants for the formation of transient resulting from the reaction of N3 • with 3-MeQ and 7-CN-3-MeQ. Interestingly, the linear plots have non-zero intercepts that indicate the involvement of equilibria (vide Discussion) and that represent first-order rate constants for the backward reactions in the equilibria.   Table 1. Table 1. Rate constants of forward and backward reactions and equilibrium constants for 3-MeQ and 7-R-3-MeQ derivatives 1 .

Kinetics Recorded in Neutral Solutions Containing 7-OCH3-3-MeQ
Experiments performed in N2O-saturated aqueous solutions containing 0.1 mM 7-OCH3-3-MeQ and 0.1 M NaN3 at pH 7 revealed that the absorption spectra were characterized by two absorption bands with maxima located at λ = 430 nm and 530 nm (vide Figure 4a). In order to directly determine the bimolecular rate constant of N3 • with 7-OCH3-3-MeQ, a kinetic analysis at various concentrations of 7-OCH3-3-MeQ (0.05 mM-0.5 mM) was performed. The growth kinetics followed at these wavelengths fit to a single exponential (Figure 7a and Figure S5). The pseudo-first-order rate constants of the formation of the 430-nm and 530-nm absorption bands were plotted as a function of 7-OCH3-3-MeQ concentration (Figure 7b). Surprisingly, it is clearly seen that the pseudo-first-order rate constants measured at both wavelengths do not show a linear dependence on the concentration of 7-OCH3-3-MeQ in the range 0.05 mM-0.5 mM. Moreover, since they do not show any specific trend and considering the errors in the measurements of absorbencies and rate constants based on The pseudo-first-order rate constants of the formation of the 430-nm and 530-nm absorption bands were plotted as a function of 7-OCH 3 -3-MeQ concentration (Figure 7b). Surprisingly, it is clearly seen that the pseudo-first-order rate constants measured at both wavelengths do not show a linear dependence on the concentration of 7-OCH 3 -3-MeQ in the range 0.05 mM-0.5 mM. Moreover, since they do not show any specific trend and considering the errors in the measurements of absorbencies and rate constants based on relatively weak signals, one can reasonably assume that they do not depend on the concentration of 7-OCH 3 -3-MeQ at all. The averaged values of the first-order rate constants are nearly equal: 3.8 × 10 5 s −1 and 4.2 × 10 5 s −1 for 430 and 530 nm, respectively (Figure 7b).
These results support the tentative hypothesis that the averaged first-order rate constant ( ∼ =4.0 × 10 5 s −1 ) can be attributed to the secondary intramolecular process leading to the transient which most likely is the 7-OCH 3 -3-MeQ •+ or its deprotonated neutral form 7-OCH 3 -3-MeQ • (vide Scheme 3 in Discussion).

Kinetics Recorded in Alkaline Solutions Containing 7-OCH 3 -3-MeQ
Experiments performed in N 2 O-saturated aqueous solutions containing 0.1 mM 7-OCH 3 -3-MeQ and 0.1 M NaN 3 at pH 11.3 revealed that the absorption spectra are characterized by two absorption bands with maxima located at λ = 430 nm and 530 nm (vide Figure 4b). Similarly, as for pH 7, in order to directly determine the bimolecular rate constant of N 3 • with 7-OCH 3 -3-MeQ (however in the anionic form), a kinetic analysis at various concentrations of 7-OCH 3 -3-MeQ (0.05 mM-0.5 mM) was performed. The rate of formation, followed at these wavelengths fits to a single exponential (Figure 8a and Figure S6). Experiments performed in N2O-saturated aqueous solutions containing 0.1 mM 7-OCH3-3-MeQ and 0.1 M NaN3 at pH 11.3 revealed that the absorption spectra are characterized by two absorption bands with maxima located at λ = 430 nm and 530 nm (vide Figure 4b). Similarly, as for pH 7, in order to directly determine the bimolecular rate constant of N3 • with 7-OCH3-3-MeQ (however in the anionic form), a kinetic analysis at various concentrations of 7-OCH3-3-MeQ (0.05 mM-0.5 mM) was performed. The rate of formation, followed at these wavelengths fits to a single exponential (Figure 8a and Figure  S6). The pseudo-first-order rate constants of the formation of the 430-nm and 530-nm absorption bands were plotted as a function of 7-OCH3-3-MeQ concentration (Figure 8b). In this case, contrary to pH 7, it is clearly seen that the pseudo-first-order rate constants measured at both wavelengths show a linear dependence on the concentration of 7-OCH3-3-MeQ in the full range of concentration studied. The slope representing second-order rate constant for the forward reaction is equal to kf = 5.5 × 10 9 M −1 s −1 calculated at both wavelengths. Both linear plots have the intercepts indicating involvement of an equilibrium (vide Scheme 3 in Discussion) and representing the first-order rate constant for the backward reactions kb = 9.7 × 10 4 s −1 and 9.0 × 10 4 s −1 calculated at 430-nm and 530-nm, respectively.

Theoretical Calculations of Radical Cations and Their Deprotonated Forms Derived from 3-MeQ and 7-R-3-MeQ Derivatives
Taking into account that SO4 •− is the strongest one-electron oxidant used in our studies one can reasonably expect that the main products of the 3-MeQ and 7-R-3-MeQ deriv- The pseudo-first-order rate constants of the formation of the 430-nm and 530-nm absorption bands were plotted as a function of 7-OCH 3 -3-MeQ concentration (Figure 8b). In this case, contrary to pH 7, it is clearly seen that the pseudo-first-order rate constants measured at both wavelengths show a linear dependence on the concentration of 7-OCH 3 -3-MeQ in the full range of concentration studied. The slope representing second-order rate constant for the forward reaction is equal to k f = 5.5 × 10 9 M −1 s −1 calculated at both wavelengths. Both linear plots have the intercepts indicating involvement of an equilibrium (vide Scheme 3 in Discussion) and representing the first-order rate constant for the backward reactions k b = 9.7 × 10 4 s −1 and 9.0 × 10 4 s −1 calculated at 430-nm and 530-nm, respectively.

Theoretical Calculations of N3 • Adducts to 3-MeQ and 7-R-3-MeQ Derivatives
In order to confirm the structure of the potential N3 • adducts to 3-MeQ and 7-R-3-MeQ derivatives, we considered two potential sites of N3 • addition, namely addition at the C-2 and C-3 carbon atoms in pyrazine moiety (vide Figure S8). The calculated UV-vis spectra at the DFT level for the respective adducts in 3-MeQ are presented in Figure 10. A reasonably good agreement is observed between the experimental spectrum and the calculated spectrum of the adduct at the C2 carbon atom, particularly regarding a distinct absorption band with λ max = 355 nm. An additional band with λ max = 540 nm is almost invisible in the experimental spectrum, which is probably related to its very low intensity being at the limit of detection of the experimental system. This observation means that addition of N 3 • to the double bond between the N1 nitrogen atom and the C2 carbon atom in pyrazine ring can only occur in the enolic tautomeric form of 3-MeQ (vide right side of the equilibrium in Figure 11). It has to be noted that the spectral features of the calculated spectrum of the adduct at the C3 carbon atom are not present in the experimental spectrum. For unknown reasons, addition of N 3 • to the double bond between the C3 carbon atom and the N4 nitrogen atom in pyrazine ring does not seem to take place. A reasonably good agreement is observed between the experimental spectrum and the calculated spectrum of the adduct at the C2 carbon atom, particularly regarding a distinct absorption band with λmax = 355 nm. An additional band with λmax = 540 nm is almost invisible in the experimental spectrum, which is probably related to its very low intensity being at the limit of detection of the experimental system. This observation means that addition of N3 • to the double bond between the N1 nitrogen atom and the C2 carbon atom in pyrazine ring can only occur in the enolic tautomeric form of 3-MeQ (vide right side of the equilibrium in Figure 11). It has to be noted that the spectral features of the calculated spectrum of the adduct at the C3 carbon atom are not present in the experimental spectrum. For unknown reasons, addition of N3 • to the double bond between the C3 carbon atom and the N4 nitrogen atom in pyrazine ring does not seem to take place. An additional proof that addition of N3 • can only occur at the C2 carbon atom in the enolic tautomeric form of 3-MeQ was obtained by performing experiments with 1,3-dimethyl-quinoxalin-2-one (1,3-diMeQ) where tautomeric equilibrium is not possible ( Figure  11). In this chemical system no absorption band was observed in the 300-700 nm region An additional proof that addition of N 3 • can only occur at the C2 carbon atom in the enolic tautomeric form of 3-MeQ was obtained by performing experiments with 1,3-dimethyl-quinoxalin-2-one (1,3-diMeQ) where tautomeric equilibrium is not possible (Figure 11). In this chemical system no absorption band was observed in the 300-700 nm region which is a strong premise that addition at the C2 carbon atom is crucial for N 3 • . Interestingly, even in this case, addition of N 3 • at the C3 carbon atom was not observed.

Reaction Pathways Involving One-Electron Oxidants and 3-MeQ/7-R-3-MeQ Derivatives
Transient absorption spectra observed on reaction of SO 4 •− with 3-MeQ (R = -H) and 7-R-3-MeQ derivatives containing electron-withdrawing substituents (−CN, −F, −CF 3 ) are characterized by the similar features, i.e., two absorption bands with λ max = 390 and 480 nm (vide Figures 2 and 3a). It means that the character of electron-withdrawing substituents does not have an influence on the location of maxima of absorption bands. On the contrary, the absorption spectrum observed on reaction of SO 4 •− with 7-OCH 3 -3-MeQ, though is characterized by two similar absorption bands, however, with maxima shifted towards longer wavelengths, λ max = 420 and 540 nm (vide Figure 4a). The similar spectrum was observed on reaction of two other one-electron oxidants, i.e., Tl 2+ and CO 3 •− (vide Figure 4a,b). Since SO 4 •− , Tl 2+ , and CO 3 •− are considered as one-electron oxidants with high or reasonably high reduction potentials (vide Table S1), these spectra could in principle be assigned to either 7-R-3-MeQ •+ and/or their deprotonated forms (7-R-3-MeQ • ). These radicals can be formed by a direct electron transfer (outer-sphere electron transfer) (Scheme 1). Generally, the pKa values of radicals are lower than the pKa values of the compounds they are derived from. The ∆pKa reflects the increase in acidity on one-electron oxidation. Such increase in acidity (with the 12 orders of magnitude) was observed for phenol (PhOH) (pKa = 10.0) and phenol radical cation (PhOH •+ ) (pKa = −2) [78] and, however smaller, for guanosine (G) (pKa = 9.4) and guanosine radical cation (G •+ ) (pKa = 3.9) [79].
The pKa values of the acid-base equilibria of indolyl radical cations (InH •+ ) and N-centered indolyl radicals (In • ) derived from indoles and tryptophan are another relevant examples. These values are located between 4.3 and 6.1 [10,11,80], and are much lower than the pKa value of indole equal to 16.97 [81]. Therefore, there is no doubt that the transient formed by a direct electron transfer between CO3 •− with 7-OCH3-3-MeQ (present in deprotonated form at pH 11.3) is N1-centered radical, 7-OCH3-3-MeQ • (Scheme 1). The question that arises at this point concerns the character of the transients which are formed at pH below pKa of 7-R-3-MeQ compounds, where they exist in protonated forms (vide Table S2). To answer this question, we compared the spectra observed on reaction of SO4 •− with 7-OCH3-3-MeQ at pH 4 and pH 7 (vide Figure S7) with the spectrum observed on reaction of CO3 •− with 7-OCH3-3-MeQ at pH 11.3 (vide Figure 4b). The same position of the absorption maxima in all three spectra is the first indication that 7-OCH3-3-MeQ • is also present at pH 4 and 7. This radical is formed by deprotonation of 7-OCH3-3-MeQ •+ (Scheme Generally, the pK a values of radicals are lower than the pK a values of the compounds they are derived from. The ∆pK a reflects the increase in acidity on one-electron oxidation. Such increase in acidity (with the 12 orders of magnitude) was observed for phenol (PhOH) (pKa = 10.0) and phenol radical cation (PhOH •+ ) (pK a = −2) [78] and, however smaller, for guanosine (G) (pK a = 9.4) and guanosine radical cation (G •+ ) (pK a = 3.9) [79]. The pK a values of the acid-base equilibria of indolyl radical cations (InH •+ ) and N-centered indolyl radicals (In • ) derived from indoles and tryptophan are another relevant examples. These values are located between 4.3 and 6.1 [10,11,80], and are much lower than the pK a value of indole equal to 16.97 [81]. Therefore, there is no doubt that the transient formed by a direct electron transfer between CO 3 •− with 7-OCH 3 -3-MeQ (present in deprotonated form at pH 11.3) is N1-centered radical, 7-OCH 3 -3-MeQ • (Scheme 1). The question that arises at this point concerns the character of the transients which are formed at pH below pK a of 7-R-3-MeQ compounds, where they exist in protonated forms (vide Table S2). To answer this question, we compared the spectra observed on reaction of SO 4 •− with 7-OCH 3 -3-MeQ at pH 4 and pH 7 (vide Figure S7) with the spectrum observed on reaction of CO 3 •− with 7-OCH 3 -3-MeQ at pH 11.3 (vide Figure 4b). The same position of the absorption maxima in all three spectra is the first indication that 7-OCH 3 -3-MeQ • is also present at pH 4 and 7. This radical is formed by deprotonation of 7-OCH 3 -3-MeQ •+ (Scheme 1). Generally, the radical cations absorb at longer wavelengths in comparison to their deprotonated forms. Again, spectra of indolyl radical cations and N-centered indolyl radicals derived from indoles and tryptophan are relevant examples [11,82]. This trend was also confirmed by the ωB97XD/aug-cc-pVTZ calculated UV-Vis spectra of the 7-OCH 3 -3-MeQ •+ and 7-OCH 3 -3-MeQ • species (vide Figure S7). With these results and premises in hands, the molar absorption coefficients of transients at the respective maxima of spectra (λ = 420 and 540 nm) were calculated to be ε 420 = 4000 M −1 cm −1 and ε 420 = 4000 M −1 cm −1 for pH 4 and 7 taking the G-value of (SO 4 •− ) = 0.29 µM J −1 and ε 420 = 4100 M −1 cm −1 and ε 420 = 3900 M −1 cm −1 for pH 11.3 taking the G-value of (CO 3 •− ) = 0.61 µM J −1 . The G-values of SO 4 •− and CO 3 •− for the respective concentrations of S 2 O 8 2− and CO 3 2− were calculated from the Schuler formula which allows correction of the G-values of their precursors (e − aq , HO • ) resulting from competition between their scavenging reactions and track recombination processes [83,84]. These calculations confirm ultimately that 7-OCH 3 -3-MeQ • is present at pH 4 and 7 which places the pK a value of the acid-base equilibrium of 7-OCH 3 -3-MeQ •+ and 7-OCH 3 -3-MeQ • below 4. This is consistent with the results obtained for guanosine (G), taking into account similar pK a values of G and 7-OCH 3 -3-MeQ in the native state.

Reaction Pathway Involving N 3 • and 3-MeQ/7-R-3-MeQ Derivatives with Electron Withdrawing Substituents (R)
Experimental observations that reactions of N 3 • with 3-MeQ and 7-R-3-MeQ (R = −CN, −CF 3 and −F) lead to significantly different absorption spectra (vide Figures 2 and 3b) compared to the absorption spectra obtained in reactions with SO 4 •− as one-electron oxidant (vide Figures 2 and 3a) clearly indicate that electron transfer does not occur. This is due to the fact that N 3 • is too weak oxidant be able to oxidize 3-MeQ and its derivatives with electron withdrawing substituents. In another words, the reduction potentials of these compounds must be higher than +1. 33 Figures 2 and 3b)? The molecular structure of the respective tautomers, the lack of a similar absorption band in the presence of 1,3-dimetylquinoxalin-2-one and the comparison of calculated absorption bands of the potential N 3 • adducts with the experimental spectra clearly show that the primary N 3 • attack occurs only on a double C=N bond at the C2 carbon atom in the enolic tautomer. This observation suggests regioselectivity in the N 3 • addition and in a consequence formation of N-centered radical on the N1 nitrogen atom (Scheme 2). These reactions occur with high-rate constants which ranged from 6.1 × 10 9 M −1 s −1 to 9.8 × 10 9 M −1 s −1 suggesting that they are nearly diffusion-controlled. Moreover, the interesting finding is reversibility of these reactions with an involvement of an equilibrium with K eq ranged from 2.9 × 10 3 M −1 to 1.2 × 10 4 M −1 (Table 1).

Reaction Pathway Involving N3 • and 3-MeQ/7-R-3-MeQ Derivatives with Electron Withdrawing Substituents (R)
Experimental observations that reactions of N3 • with 3-MeQ and 7-R-3-MeQ (R = −CN, −CF3 and −F) lead to significantly different absorption spectra (vide Figure 2 and Figure 3b) compared to the absorption spectra obtained in reactions with SO4 •− as oneelectron oxidant (vide Figure 2 and Figure 3a) clearly indicate that electron transfer does not occur. This is due to the fact that N3 • is too weak oxidant be able to oxidize 3-MeQ and its derivatives with electron withdrawing substituents. In another words, the reduction potentials of these compounds must be higher than +1. 33 Figure 3b)? The molecular structure of the respective tautomers, the lack of a similar absorption band in the presence of 1,3dimetylquinoxalin-2-one and the comparison of calculated absorption bands of the potential N3 • adducts with the experimental spectra clearly show that the primary N3 • attack occurs only on a double C=N bond at the C2 carbon atom in the enolic tautomer. This observation suggests regioselectivity in the N3 • addition and in a consequence formation of N-centered radical on the N1 nitrogen atom (Scheme 2). These reactions occur with high-rate constants which ranged from 6.1 × 10 9 M −1 s −1 to 9.8 × 10 9 M −1 s −1 suggesting that they are nearly diffusion-controlled. Moreover, the interesting finding is reversibility of these reactions with an involvement of an equilibrium with Keq ranged from 2.9 × 10 3 M −1 to 1.2 × 10 4 M −1 (Table 1).

Reaction Pathways Involving N3 • and 7-OCH3-3-MeQ
In the case of 7-OCH3-3-MeQ, the absorption spectrum of the transient formed by reaction with N3 • at pH 7 is similar to absorption spectra formed by reaction with SO4 •− and Tl 2+ (vide Figure 4a). This spectrum is characterized by similar features in comparison

Reaction Pathways Involving N 3 • and 7-OCH 3 -3-MeQ
In the case of 7-OCH 3 -3-MeQ, the absorption spectrum of the transient formed by reaction with N 3 • at pH 7 is similar to absorption spectra formed by reaction with SO 4 •− and Tl 2+ (vide Figure 4a). This spectrum is characterized by similar features in comparison to absorption spectra observed for the other 7-R-3-MeQ derivatives, except for a small shift of two absorption maxima towards longer wavelengths (compare Figures 3a and 4a).
On the basis of the previous assignments for 7-R-3-MeQ derivatives and the ωB97XD/augcc-pVTZ calculated UV-Vis spectra (vide Figure S7), the spectra presented on Figure 4a were assigned to 7-OCH 3 -3-MeQ • . Furthermore, absorption spectrum with similar features and consisted of two absorption bands with maxima located at λ = 460 and 560 nm was assigned to 7-NH 2 -3-MeQ • formed by reaction of N 3 • with 7-NH 2 -3-MeQ (vide Figure S3). These observations clearly indicate that N 3 • are able to oxidize 7-OCH 3 -3-MeQ and 7-NH 2 -3-MeQ and impose values of their reduction potentials lower than +1.33 V vs. NHE. This is in line with the expected decrease of reduction potentials of 7-R-3-MeQ derivatives with electron-donating substituents.
Lack of the expected linear dependence of the pseudo-first-order rate constants of the formation of 7-OCH 3 -3-MeQ • on concentration of 7-OCH 3 -3-MeQ at pH 7 requires explanation. The following mechanism is proposed to explain the experimental observations (Scheme 3). The initial attack of N3 • , first step, is by addition to the double bond C=N at the C2 carbon atom in the enolic tautomer producing N-centered radical on the N1 nitrogen atom. This radical undergo elimination of N3 − anion ("inner-sphere" electron transfer) leading to the 7-OCH3-3-MeQ •+ (second step) which further undergoes deprotonation leading to the 7-OCH3-3-MeQ • (third step). Since the rate of formation of 7-OCH3-3-MeQ • does not depend on the concentration of 7-OCH3-3-MeQ, the first step does not control its formation. The averaged value of the first-rate constant (k ≅ 4.0 × 10 5 s −1 ) measured in the full range of concentration studied represents rather the rate of the second step which does not depend on 7-OCH3-3-MeQ concentration.
Moreover, it is interesting to note that at pH 11.3 the pseudo-first-order rate constants of the formation of 7-OCH3-3-MeQ • depend linearly on concentration of 7-OCH3-3-MeQ (vide Figure 8b). The second order rate constant kf = 5.5 × 10 9 M −1 s −1 and the first order rate constants kb = 9.7 × 10 4 s −1 and 9.0 × 10 4 s −1 represent the respective rate constants of forward and backward reactions in the equilibrium displayed at the bottom of Scheme 3. These The initial attack of N 3 • , first step, is by addition to the double bond C=N at the C2 carbon atom in the enolic tautomer producing N-centered radical on the N1 nitrogen atom. This radical undergo elimination of N 3 − anion ("inner-sphere" electron transfer) leading to the 7-OCH 3 -3-MeQ •+ (second step) which further undergoes deprotonation leading to the 7-OCH 3 -3-MeQ • (third step). Since the rate of formation of 7-OCH 3 -3-MeQ • does not depend on the concentration of 7-OCH 3 -3-MeQ, the first step does not control its formation. The averaged value of the first-rate constant (k ∼ = 4.0 × 10 5 s −1 ) measured in the full range of concentration studied represents rather the rate of the second step which does not depend on 7-OCH 3 -3-MeQ concentration.

Preparation of Solutions
All solutions were made with water triply distilled provided by a Millipore Direct-Q 3-UV system. The pH was adjusted by the addition of NaOH or HClO 4 . Prior to irradiation, the samples were purged gently with N 2 O for 30 min. per 200 mL volume. The typical concentration of 7-substituted 3-methyl-2(1H)-quinoxalin-2-ones in solutions was 0.1 mM, unless otherwise specified. Experiments were performed with a continuous flow of sample solutions using a standard quartz cell with optical length 1 cm at room temperature (~23 • C).

Pulse Radiolysis Instrumentation
The pulse radiolysis experiments were performed with the LAE-10 linear accelerator at the Institute of Nuclear Chemistry and Technology in Warsaw, Poland with typical electron pulse length of 8 ns and 10 MeV of energy. A detailed description of the experimental setup has been given elsewhere along with basic details of the equipment and its data collection system [86,87]. The 150W xenon arc lamp E7536 (Hamamatsu Photonics K.K) and 1 kW UV-enhanced xenon arc lamp (Oriel Instruments) were alternately applied as monitoring light sources. The respective wavelengths were selected by MSH 301 monochromator (Lot Oriel Gruppe) with resolution 2.4 nm. The intensity of analyzing light was measured by means of PMT R955 (Hamamatsu). A signal from detector was digitized using a Le Croy WaveSurfer 104MXs-B (1 GHz, 10 GS/s) oscilloscope and then send to PC for further processing. Water filter was used to eliminate near IR wavelengths.

Dosimetry
The dosimetry was based on N 2 O-saturated solutions of 10 −2 M KSCN which, following radiolysis, produces (SCN) 2 •− radicals that have a molar absorption coefficient of 7580 M −1 cm −1 at λ = 472 nm and are produced with a yield of G = 0.635 µmol J −1 [88]. Absorbed doses per pulse were on the order of 20 Gy (1 Gy = 1 J kg −1 ).
In turn, in Ar-saturated aqueous solutions, HO • radicals can be selectively removed by the addition of 2-methyl-2-propanol (tert-butanol) according to Equation (10)

Theoretical Procedures
All DFT [93] or TD-DFT [94] calculations, were they restricted or unrestricted, were performed using the dispersion correction ωB97XD functional [95] combined with the Dunning's aug-cc-pVTZ correlation-consistent polarized basis sets augmented with diffuse functions [96,97] using the Gaussian 09 revision D.01 suite of programs [98]. The ωB97XD functional was recommended for calculation of many different properties [99][100][101][102], and the aug-cc-pVTZ basis set was among the most successful and widely used basis sets for post-Hartree-Fock and diverse DFT studies [101,[103][104][105][106]. To ascertain that the optimized structures were true minima, the harmonic frequencies of all of them in the ground, radical and excited states were determined to be real. The charge, spin, and population analysis was conducted according to the NBO method [107] as implemented in the Gaussian 09 revision D.01 suite of programs [98]. Correlation analysis was done using the SigmaPlot 13 program [108] (version 13). The TD-DFT methods are known to predict spectra that are shifted from the experimental ones due to the basis set incompleteness and DFT functional inadequacy. Moreover, different transitions are often reproduced with a different shift. These factors always play a role especially when absorption spectra of transient radical species are calculated using a single reference method. Here, we manually shifted the calculated spectra by 20-130 nm to lower energies to best match the experimental ones.

Conclusions
In the current paper, we provided an experimental approach that allows to clearly define the nature of the obtained products formed in reactions of N 3 • with quinoxalin-2ones in aqueous solutions. This approach is based on comparison of transient absorption spectra observed on reaction of N 3 • with those observed on reaction of one-electron oxidants (e.g., SO 4 •− , Tl 2+ and CO 3 •− ) with 3-MeQ and a series of 7-R-3-MeQ derivatives with electron withdrawing and electron-donating substituents. The mere fact that reactions of N 3 • with quinoxalin-2-ones take place with high-rate constants (> 10 9 M −1 s −1 ) does not ultimately prove that they are electron transfer reactions. For 3-MeQ and 7-R-3-MeQ derivatives with electron withdrawing substituents, the absorption spectra formed in reactions involving N 3 • were different from the absorption spectra formed in the reactions involving other one-electron oxidants reacting only by electron transfer. Based on calculated absorption spectra employing density functional theory (DFT and TD-DFT), these spectra were assigned to N 3 • adducts at the C2 carbon atom in pyrazine moiety. On the other hand, for 7-OCH 3 -3-MeQ the absorption spectra formed in reactions involving N 3 • and other one-electron oxidants were similar and assigned to 7-OCH 3 -3-MeQ • . Interestingly, depending on pH, formation of 7-OCH 3 -3-MeQ • can occur by either addition of N 3 • to the pyrazine ring followed by elimination of SO 4 •− ("inner-sphere" electron transfer) or by a direct electron transfer ("outer-sphere" electron transfer), and followed by deprotonation of 7-OCH 3 -3-MeQ •+ . The approach presented here can be applied to other organic molecules containing double bonds for the proper assignment of absorption spectra either to N 3 • one-electron oxidation products or its addition products. Such knowledge is important for evaluating the usefulness of N 3 • as a secondary oxidant in biological studies in aqueous phase.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
All data is displayed in the manuscript.