Radiation-Driven Formation of Reactive Oxygen Species in Oxychlorine-Containing Mars Surface Analogues

2.1. Reagents

Acetonitrile (ACN, HPLC grade cat. # 34998), hydrogen peroxide solution (H₂O₂ 30%, cat. # H1009), xylenol orange (XO, Na₄-salt, cat. # X3500), ferrous ammonium sulfate hexahydrate (FAS, cat. # 215406), catalase (CAT, cat. # C3155), dimethyl sulfoxide (DMSO, cat. # 472301), potassium superoxide (KO₂, cat. # P0647), sodium chlorite (NaClO₂, cat. # 244155), magnesium peroxide (MgO₂, cat. # 63130), sodium peroxide (Na₂O₂, cat. # 311456), and hydroethidine (HE, cat. # D7008) were obtained from Sigma-Aldrich (St. Louis, MO, USA). D-Sorbitol (anhydrous, cat. # 194742) was obtained from MPI Biochemicals (Santa Ana, CA, USA). Calcium peroxide (CaO₂, cat. # 21157), dicyclohexano-18-crown-6 ether (CE, cat. # A15344), ascorbic acid (cat. # 36237), and terephthalic acid (TPA, cat. # 43423) were obtained from Alfa Aesar (Karlsruhe, Germany). Sodium hypochlorite solution (NaClO, cat. # 105614) and CuSO₄·5H₂O (cat. # 102787) were obtained from Merk (Darmstadt, Germany). All other chemicals used in this study were reagent grade.

2.2. Preparation of perchlorate samples

The Mars Phoenix salt analogue is chemically analogous to the 80 mg water-soluble salts detected per 1 cc of martian soil and is described by Quinn et al. (2011). This analog mixture includes 58.3 μmol ClO₄ ⁻ and is referred to as the complete analogue. A second version of this analogue, which does not contain perchlorate, is referred to as the incomplete salt analogue (Quinn et al., 2011). Gamma irradiation (500 kGy) was performed as described by Quinn et al. (2013), using samples sealed in glass ampules under 7.5 mbar CO₂ to simulate the martian atmosphere (Table 1).

2.3. Inorganic-origin superoxide (O₂ ^•−) and peroxide (H₂O₂) assay

A new assay for the simultaneous determination of O₂ ^•− and H₂O₂ generated from the Mars Phoenix analogues was developed to study the possible γ-radiation production of these species. This assay was developed by using KO₂ and was experimentally validated with MgO₂, CaO₂, and Na₂O₂ (Table 2), and NaClO and NaClO₂ (Table 3). The detailed methods used in these experiments are presented in the Supplementary Materials (available online at www.liebertonline.com/ast). In the present section, the assay principle and its validation experiments are briefly overviewed.

Assay principle and validation: The superoxide-peroxide assay is an extension of the Fe²⁺-xylenol orange (FOX) assay (Grintzalis et al., 2013; Georgiou et al., 2015) for soil O₂ ^•−. It measures sample O₂ ^•− (after dismutation to H₂O₂ and O₂) and sample H₂O₂ in separate assay steps following extraction in alkaline acetonitrile (ACN_alk) containing dicyclohexano-18-crown-6 ether (CE). After extraction, the ACN_alk-CE extract is split into two fractions, one for O₂ ^•− and the other for H₂O₂ determination. In the O₂ ^•− determination procedure, CAT is used to eliminate H₂O₂, and subsequently O₂ ^•− is dismutated using ddH₂O (at 50% ACN final concentration) and quantified as H₂O₂. Sample O₂ ^•− concentration is determined based on O₂ ^•− dismutation stoichiometry (O₂ ^•− + H₂O → OH⁻ + ½H₂O₂ + ½O₂). For H₂O₂ determination, CAT is not used, and O₂ ^•− is dismutated (at 50% ACN in water) to H₂O₂. Total H₂O₂ is determined, and then the concentration of sample H₂O₂ is determined by subtracting the measured quantity H₂O₂ from the quantity of H₂O₂ measured in the O₂ ^•− determination procedure.

The superoxide-peroxide assay was validated for the quantification of O₂ ^•− by using two different standard methods for O₂ ^•− determination. In the first method, a HE-based fluorometric assay (Georgiou et al., 2015), O₂ ^•− (from reagent KO₂) was measured (at 1 mM HCl, 40 μ M HE) in runs by using ACN (containing 0.2–1 mM CE) over a concentration range of 95–20% (at pH 7.0) with 99% ACN_alk as control. The second method, the superoxide dismutase–inhibited reduction of cytochrome c by O₂ ^•−, was used to test the dismutation of O₂ ^•− (from reagent KO₂) in ACN over a concentration range up to 50%. The results of these tests were compared to the method developed for the present study. All three methods gave same results on the O₂ ^•− measurement in the tested 95–20% ACN solutions, and show that the dismutation of O₂ ^•− radical takes place below 80% ACN (Fig. 1a). Moreover, the tested incubation intervals with CAT (500 and 50 units) of the (respective 95–70% and 60–20% range) ACN O₂ ^•− solutions in the second method showed that O₂ ^•− was stable even at the limit of the 80% ACN concentration and for the 10 min incubation period used in the superoxide-peroxide assay.

OPEN IN VIEWER

FIG. 1.

Parameters of the inorganic origin O₂ ^•−/H₂O₂ assay: (a) Dismutation of O₂ ^•− (21 μM) as a function of ACN concentration, measured directly (as O₂ ^•−) and indirectly as H₂O₂ (the product of O₂ ^•− dismutation) by the HE-based fluorometric (direct) in comparison to the present assay (indirect). (b) H₂O₂-based slope of the assay, expressed as absorbance (at 560 nm) per 1 μM H₂O₂, as a function of ACN concentration. (c₁ ) Decomposition of H₂O₂ (100% = 1 mM) by CAT (from stock 20 KU/mL) in 95% ACN (in 15 mM phosphate buffer, pH 7.5, and 0–5 mM CE) after incubation for 10 min. The assay was tested in 500-fold dilutions of the incubation mixture (at ∼0% ACN). (c₂ ) Decomposition of H₂O₂ (100% = 100 μ M) by CAT in 50% ACN (in 10 mM phosphate buffer, pH 7.5) after incubation for 5 min. The assay was tested in 50-fold dilutions of the incubation mixture (at ∼50% ACN). Inset shows CAT interference at 50% ACN in the assay as measured by the assay absorbance at 560 nm. Error bars designate standard error.

The effect of ACN-CE concentration on the superoxide-peroxide assay standard curve (0–2 μ M H₂O₂) was tested to determine the optimum assay sensitivity (i.e., the maximum standard curve slope for different ACN_alk-CE extract concentrations). The slope is an exponential function of % ACN (Fig. 1b), with CE concentration not affecting the slope (tested to 8 mM). Based on the exponential function, the maximum slope to sample dilution factor ratio (0.13/2) is achieved at 50%. (For example, at 20% the slope to sample dilution factor is 0.17/5. Therefore, the optimum ACN concentration was chosen at 50% for performing the assay.)

The CAT activity in various concentrations of ACN was examined to verify its efficiency to decompose H₂O₂ over a range of incubation intervals (up to 10 min) and H₂O₂ concentrations (up to 2 mM). The results of this test verified that co-extracted O₂ ^2- is removed from sample fraction 1, without loss of O₂ ^•− (Fig. 1c ₁). Since CAT heme iron may interfere with the assay, the optimum CAT units for the decomposition of H₂O₂ (at assay final 50% ACN, containing 10 mM phosphate buffer, pH 7.5) were determined. As control, CAT interference (at 50% ACN) was measured (as absorbance at 560 nm) in the absence of H₂O₂. Based on these results to minimize its heme iron interference, a maximum 30 units CAT (per milliliter assay mixture) was used in the assay (Fig. 1c ₂).

Assay sensitivity was determined to be ∼70 pmol O₂ ^•− and 35 pmol H₂O₂ co-extracted in 100% ACN (and determined at 50% ACN) and 14 pmol H₂O₂ when the assay is used only for H₂O₂ determination in H₂O extracts of solid samples. The details of assay protocol used for peroxide and superoxide determinations are given in the Supplementary Materials.

2.4. Hydroxyl radical assay

Quantification of ^•OH is performed by a modification of a previous method (Georgiou et al., 2015). The method is based on the specific reaction of ^•OH with TPA in Na-borate buffer, pH 9, and the formation of the fluorescent product 2-hydroxy terephthalic acid (2HTPA) measured at 423 nm (311 nm excitation). Net peak emission fluorescence is obtained against the controls Na-borate-TPA and Na-borate-sample (at the highest tested concentration). Identification of ^•OH is performed by the inhibition of 2HTPA formation with the presence in a Na-borate-TPA assay solution of 100 mM DMSO, a well-known scavenger of ^•OH. The net fluorescence units (FU) of 2HTPA are converted to nanomoles of ^•OH by a standard curve (see the Supplementary Materials).

2.5. Effect of ClO₂ ⁻ and ClO⁻ on ^•OH standard curve

Since production levels of ^•OH are assessed indirectly by the production levels of 2HTPA, its quantification by the ^•OH assay may be underestimated due to its fluorescence quenching by the presence of possible oxidants such as ClO⁻ and ClO₂ ⁻. These are well-known products of the radiolysis of metal perchlorate (Prince and Johnson, 1965; Paviet-Hartmann et al., 2001; Quinn et al., 2013); thus their possible presence in radiolyzed Mg(ClO₄)₂ alone/mixed with the Mars Phoenix site salt analogue may underestimate 2HTPA (thus ^•OH) production levels. The same 2HTPA oxidation issue will arise when assessing photolyzed ClO⁻ and ClO₂ ⁻ as potential sources of ^•OH. Another consideration arises from the fact that the actual molar ratio 2HTPA/^•OH (the theoretical is 1:1) must be determined under the conditions of the ^•OH assay in order to estimate accurately the production levels of ^•OH. This is established by the construction of a standard curve of 2HTPA FU versus ^•OH under the same conditions for the ^•OH assay. Details of these tests are given in the Supplementary Materials.

2.6. NaClO solidification and heating: recovery

For investigating ClO⁻ as generator of H₂O₂ and ^•OH (Table 3 and Fig. 4, respectively) in the solid form, as produced in radiolyzed Mg(ClO₄)₂, “solid” ClO⁻ was formed by drying (evaporating) NaClO mixed with fumed silica. Recovery was determined after solidification and heating at 160°C for 3 h. It should be noted that NaClO solution upon drying produces—stable at ∼20°C to ∼50°C—the hydrated forms NaClO·5H₂O and NaClO·2.5H₂O (Elsenousy and Chevrier, 2014). Details of these tests are given in the Supplementary Materials.

2.7. Products from UVC-photolyzed ClO⁻ and ClO₂ ⁻

ClO⁻ and ClO₂ ⁻ were photolyzed by UVC in order to study, under simulated martian topsoil UV exposure conditions, their conversion to Cl-based products, which could act as photolysis-induced generators of ^•OH (see subsequent experiment and Fig. 4). Experimental details are given in the Supplementary Materials.

2.8. Generation of ^•OH by radiolyzed Mg(ClO₄)₂ and the Mars Phoenix site salt analogue, and by metal peroxide analogues

Mg(ClO₄)₂, the Mars Phoenix salt analogue (with and without perchlorate; radiolyzed and nonradiolyzed), and commercial MgO₂, CaO₂, and Na₂O₂ were tested for ^•OH production (Tables 1 and 2), following the procedure of the aforementioned ^•OH assay.

2.9. ^•OH generation by UVA photolysis of ClO⁻ and ClO₂ ⁻

Due to the very fast UVC photolysis of NaClO and NaClO₂ (Fig. 3), ClO⁻ and ClO₂ ⁻ were investigated as potential sources of ^•OH under time-controlled UVA exposure (Fig. 4). Experimental details are provided in the Supplementary Materials.

2.10. Statistical analysis

All data are reported as mean ± standard error of at least three independent experiments and were analyzed with the SPSS statistical package (SPSS Inc., 2001, Release 11.0.0, USA). Whenever appropriate, the significance was determined for Student's unpaired t test or ANOVA. A value of P < 0.05 was considered to be significant.

3.1. Superoxide and peroxide production by radiolysis of Mg(ClO₄)₂ and Phoenix salt analogue

Generation of O₂ ^•− and H₂O₂ by radiolysis of Mg(ClO₄)₂ and the Phoenix salt analogues is shown in Table 1. O₂ ^•− was detected only in the perchlorate-containing salt analogue at 1 nmol per 80 mg of analog sample (80 mg of salt sample corresponds to the measured salt content in 1 cc of regolith at the Phoenix site and contains 58.3 μmol ClO₄ ⁻). The same sample produced ∼2 nmol of H₂O₂. No O₂ ^•− and 3.3-fold less H₂O₂ were measured in the sample after heating at 160°C for 3 h. H₂O₂ was also produced, at a level of ∼6 nmol per 58.3 μmol ClO₄ ⁻, by radiolyzing the pure Mg(ClO₄)₂, and was reduced by 61-fold by heating the sample. No H₂O₂ or O₂ ^•− was measured in the radiolyzed salt mixture that did not contain perchlorate. In addition, no H₂O₂ or O₂ ^•− production was measured in any of the three sample types if they were not irradiated (control samples).

Control experiments were run on metal peroxides MgO₂, CaO₂, and Na₂O₂ to verify the ability of the assay to quantify O₂ ^•− and H₂O₂ and to determine whether they may be sources of H₂O₂ via hydrolysis (Table 2) if they are generated during radiolysis of Mg(ClO₄)₂ and/or the Phoenix site salt analogue (with and without perchlorate). Additionally, NaClO and NaClO₂ were tested using the assay to determine whether they generate H₂O₂ (Table 3) because hypochlorite (ClO⁻) and chlorite (ClO₂ ⁻) are produced by radiolyzed perchlorate (Prince and Johnson, 1965; Paviet-Hartmann et al., 2001; Quinn et al., 2013).

H₂O₂ was measured in MgO₂ and Na₂O₂ (unheated) ACN_alk-CE extracts at levels of 0.029 and 0.064 mol %, respectively, which are comparable to the levels measured with radiolyzed Mg(ClO₄)₂ (0.011 mol %). H₂O₂ levels measured in CaO₂ (unheated) ACN_alk-CE extracts were 2.5- and 5.5-fold higher than those measured for Na₂O₂ and MgO₂, respectively. H₂O₂ generation by MgO₂, CaO₂, and Na₂O₂ was decreased by heating them (at 160°C for 3 h) by 1.8-fold on average, while heating decreased H₂O₂ generation by radiolyzed Mg(ClO₄)₂ by 61-fold (Tables 1 and 2). The ACN-insoluble fractions of unheated and heated MgO₂, CaO₂, and Na₂O₂ produced much higher levels of H₂O₂ upon solubilization in ddH₂O than the soluble fractions, with ddH₂O-dissolved MgO₂ producing the lowest levels (0.14 and 0.056 mol % unheated and heated, respectively; Table 2).

On the other hand, the ACN_alk-CE-extracted unheated and heated (at 160°C for 3 h) controls NaClO₂ and NaClO produced ∼80- and ∼150-fold higher H₂O₂ concentration levels (0.9 and 1.7% mol) than ACN_alk-CE-extracted unheated radiolyzed Mg(ClO₄)₂, respectively (Tables 3 and 1). This is consistent with the fact that the levels of ClO₂ ⁻ and ClO⁻ in the radiolyzed Mg(ClO₄)₂ are only a small mole fraction of the sample. The concentration (% mol) of H₂O₂ as the sum of the ACN- and ACN/H₂O-solid extract fractions for the unheated and heated NaClO (∼3) is almost 40% of that for the non-ACN-fractionated “liquid unheated,” “dry unheated,” or “dry heated” controls (Table 3), which suggests a less than 100% recovery during the ACN fractionation procedure.

The concentration of H₂O₂ produced by heating NaClO was determined in comparison to the fraction that was retained after heating. NaClO was solidified in fumed silica, and its recovery was measured after heating to test it as a solid-state generator of ^•OH upon UV exposure. Liquid NaClO was solidified in fumed silica by vacuum evaporation as described in Materials and Methods and in Fig. 4a ₁. Recovery of ClO⁻ by the solidification process (water removal at <50°C) was 55.6%, while after heating (for 3 h at 160°C) recovery was ∼3%. NaClO₂ was heat stable.

3.2. ^•OH generated by radiolyzed Mg(ClO₄)₂ and Mars Phoenix site salt analogue: ^•OH assay and control experiments

Radiolyzed Mg(ClO₄)₂ produced ∼1 nmol ^•OH (0.0017 mol %) per 58.3 μmol ClO₄ ⁻ when unheated. Interestingly, ∼7-fold higher levels were measured in the radiolyzed samples after they were heated. More importantly, both unheated complete and incomplete salt analogues produced a ∼150-fold higher concentration of ^•OH (149 and 147 nmol per 80 and 73.5 mg, respectively, corresponding to 1 cc martian soil). ^•OH production by the complete and the incomplete analogues decreased by an average of ∼2.5-fold when heated (Table 1). However, ^•OH production by the heated complete salt analogue (75 nmol; Table 1) was 33% higher than by the heated incomplete. The nonradiolyzed controls of all these samples did not produce ^•OH.

Hypochlorite (ClO⁻) and chlorite (ClO₂ ⁻) were investigated as possible sources of ^•OH in radiolyzed Mg(ClO₄)₂ and Mars Phoenix site salt analogue because they have been identified in the main radiolysis products of radiolyzed metal perchlorate [and Mg(ClO₄)₂] (Prince and Johnson, 1965). These were studied by using their corresponding analogues NaClO and NaClO₂ in an extensive array of experiments designed to test the following: (i) whether NaClO and NaClO₂ oxidize the fluorescent 2HTPA (Fig. 2) by which ^•OH is indirectly quantified, resulting in underestimation of ^•OH concentration; (ii) UV photolysis of NaClO and NaClO₂ with the intent to study them as possible sources of Cl-based by-products (Fig. 3) under simulated martian UVC exposure conditions; and (iii) subsequently NaClO and NaClO₂ as generators of ^•OH by a more time-controlled UV exposure (using the less energetic UVA) (Fig. 4).

OPEN IN VIEWER

FIG. 2.

Effect of ClO⁻ and ClO₂ ⁻ on the ^•OH assay standard curve: The effect of NaClO (a) and NaClO₂ (b) concentrations (nanomolar and micromolar, respectively) was studied on the fluorescence quenching of 2HTPA (the specific product of the reaction of ^•OH with its specific trap TPA). 2HTPA FU were studied as a function of (i) the concentration of NaClO and NaClO₂ [since ClO⁻ and ClO₂ ⁻ are generated by the radiolysis of perchlorate (Prince and Johnson, 1965; Paviet-Hartmann et al., 2001) and by UVC photolysis of ClO₂ ⁻ shown in Fig. 3], and (ii) their time exposure to UVA (serving also as control to their study as generators of ^•OH; see Fig. 4).

OPEN IN VIEWER

FIG. 3.

UVC photolysis products of NaClO and NaClO₂: (a) Spectra of ∼0.2 μmol NaClO (in 0.3 mL solution); inset (a₁) presents the peak absorbance change of NaClO at 292 nm versus UVC exposure time. (b) Spectra of ∼1.0 μmol NaClO₂ (in 0.3 mL solution) at zero time exposure. Actually, the zero time spectra of NaClO₂ should have only one maximum absorbance peak at 260 nm. However, they appear to contain an extra peak at 292 nm (belonging to ClO⁻ as identified by spectrum deconvolution), most likely the outcome of exposing NaClO₂ to the light source of the spectrophotometer during 250–400 nm spectrum recording [giving an apparent molar ratio 4.8 NaClO₂/NaClO, based on their molar absorption coefficients 108 and 335 M ⁻¹, respectively (Paviet-Hartmann et al., 2001)]. However, this question is resolved in the analysis of the spectra obtained after exposing NaClO₂ to UVC for 5, 10, and 15 min. The resulted spectra are composed of overlapping peaks, which after being resolved (see corresponding thinner lines) by a spectral peak deconvolution software (see Materials and Methods) were identified to belong to ClO⁻ (292 nm), Cl₃ ⁻ (325 nm), and ClO₂ ^• (360 nm) (Paviet-Hartmann et al., 2001). Panel (b₁ ) presents the micromolar change (in 0.3 mL solution) of these peaks as a function of UVC exposure time. The curve fit of the 290 nm peak absorbance value changes for ClO⁻ during 5, 10, and 15 min UVC exposure shows that its absorbance value at zero minutes' exposure is zero. The same conclusion is derived from the oxidation of 2HTPA by NaClO₂, where the curves of 2HTPA fluorescence quenching as function of different concentrations of NaClO₂ at different UVC exposure times converge to the initial FU of 2HTPA at zero time exposure (Fig. 2b). This convergence would not have happened if NaClO₂ and NaClO coexisted at a molar ratio 4.8, as NaClO would have been present at high nanomolar concentration that would have oxidized 2HTPA (decreased its initial FU) even at zero-time UVC exposure (Fig. 2a).

OPEN IN VIEWER

FIG. 4.

Hydroxyl radical generation by UVA photolysis of ClO⁻ and ClO₂ ⁻: ^•OH is shown to be generated by their commercial analogues NaClO (a) and NaClO₂ (b), respectively (unheated, solid symbols, and heated, open symbols), at different concentrations as function to UVA exposure time. Inset in (a) is a magnification of the concentration rate curve of ^•OH produced by heated NaClO, the solidification of which and its heat recovery (determined by difference spectra shown in a₁ ) are described in Materials and Methods.

(i) It was found that both NaClO and NaClO₂ oxidize 2HTPA. At 0 s UVA exposure, 7.5 and 38 nmol NaClO oxidize 2HTPA moderately. Higher concentrations oxidized to a much greater degree, resulting in an underestimation of ^•OH concentration by the employed assay. For example, 114 nmol NaClO decreased the initial 350 FU of 2HTPA to 150 FU (Fig. 2a). On the other hand, NaClO₂ did not oxidize 2HTPA even at 1000-fold higher quantities (1.6 to 24 μmol), and is shown by the extrapolation of the UVA exposure 2HTPA FU curves, which cross the y axis at the initial 350 FU (Fig. 2b). However, the oxidative destruction of the HTPA FU by NaClO₂ is evident as a function of UVA photolysis time. This is possibly due to the UVA-induced release of NaClO by photolyzed NaClO₂ as is seen in UVC photolysis (Fig. 3b), and the concomitant production of ^•OH followed by its 2HTPA oxidizing effect (Fig. 4b and unpublished experimental observation). This is also supported by the increasing destruction of the initial FU of 2HTPA (at 0 s UVA exposure) by NaClO, as the concentration of the latter increases from 7.6 to 114 nmol (Fig. 2a). On the other hand, the destruction of 2HTPA FU by the UVA photolysis of NaClO becomes more profound at high NaClO concentrations (Fig. 2a) and is possibly due to the oxidation of 2HTPA by ^•OH (as it is shown to be produced at higher rates; Fig. 4a).

(ii) Possible UVC photolysis products of NaClO and NaClO₂ were investigated spectrally under simulated martian UV exposure conditions (Fig. 3). It was found that NaClO decreased in concentration without producing any spectrally identified Cl-based species (Fig. 3a, 3a ₁). On the other hand, NaClO₂ was photolyzed to the Cl-based oxidants ClO⁻ (292 nm), Cl₃ ⁻ (325 nm), and ClO₂ ^• (360 nm). The concentration changes of these species versus exposure time were determined by their known molar extinction coefficients (Paviet-Hartmann et al., 2001). The individual spectral peaks were resolved by the spectral peak deconvolution of the composite spectra, obtained at various UVC exposure-time intervals (5, 10, 15 min).

(iii) After investigations i and ii as described above, NaClO and NaClO₂ were then tested for ^•OH generation under simulated martian soil UV exposure conditions. The aim was to uncover a possible mechanism of ^•OH generation based on the UVA photolysis of the radiolysis products of Mg(ClO₄)₂ (Fig. 4), which would complement its radiolysis mechanism for ^•OH production (shown in Table 1). Unheated and heated NaClO (i.e., solidified) and NaClO₂ both produced ^•OH only upon photolysis by UVA (311 nm) at an increasing rate as a hyperbolic function of exposure time, which became horizontal to the x axis for NaClO at low concentrations (Fig. 4a). The fact that 30 nmol of heated NaClO produced ^•OH at a 25-fold lower initial concentration rate than the average concentration of the 23 and 46 nmol unheated NaClO may be attributed to the fact that the heated sample contained NaCl in an approximate concentration 4 M in the ^•OH assay 0.3 mL solution (for explanation, see the Supplementary Materials). Such high Cl⁻ anion concentration is known to scavenge ^•OH (Liao et al., 2001). On the other hand, the rate curves for NaClO₂ peaked and then declined over extended time exposure. This could be due to the possible production of the photolysis product NaClO (Fig. 3a), which oxidizes 2HTPA (Fig. 2).

4.1. Methodological justification of the inorganic-origin superoxide/peroxide assay

Radiolyzed Mg(ClO₄)₂ and complete and incomplete Mars Phoenix site salt analogue were analyzed for generation of H₂O₂, O₂ ^•−, and ^•OH (Table 1) and compared (i) to peroxide analogues based on the metal ions present the salt analogues and (ii) known perchlorate radiolysis product analogues. The assay used for analysis extracts both O₂ ^•− and H₂O₂ from samples in an anhydrous alkaline ACN_alk-CE (or ACN_alk) solvent, which is then split into two fractions. Each fraction is then subjected to different reaction treatments in order to identify and quantify stable O₂ ^•− and H₂O₂. The first fraction contains O₂ ^•− that may be present in the sample as metal superoxides (extracted by metal cation dissociation using CE, and at the same time stabilized in the ACN_alk solvent) or trapped inside mineral lattices, which is quantified as H₂O₂ upon dismutation. In the same fraction, the assay also determines sample peroxides (free or as metal peroxides). Although H₂O₂ is normally unstable at high alkaline pH, it remains stable at the ACN_alk-CE solvent's mild alkaline pH (∼8) (Georgiou et al., 2016). Subsequent CAT treatment steps discriminate the H₂O₂ fraction attributed to O₂ ^•− from that attributed to peroxides. The second ACN_alk-CE solvent fraction also contains ACN_alk-CE-soluble metal peroxides and superoxides, which are detected as H₂O₂ after hydrolysis and dismutation (in H₂O), respectively. The ACN_alk-CE fractionation step of the present assay proved to be advantageous for the Mg(ClO₄)₂ component because it is completely solubilized in the ACN solvent, which is crucial for the identification of the possible sources of O₂ ^•− and H₂O₂.

4.2. Generation of H₂O₂ and O₂ ^•− in radiolyzed Mg(ClO₄)₂ and the Mars Phoenix salt analogue

We find that the amount of H₂O₂ measured in radiolyzed Mg(ClO₄)₂ extracted with ACN_alk-CE solvent is 3.3-fold higher (0.011%) compared to the ACN-extracted complete salt analogue (0.0033% H₂O₂). This suggests that in the salt analogue that contained Mg(ClO₄)₂, the O₂ ^•− that forms may be stabilized by the Na⁺, K⁺, and Ca²⁺ that are present in the analogue. Supporting evidence for this O₂ ^•− stabilization pathway comes from the low levels of H₂O₂ that are detected in the H₂O extracts of the complete and the incomplete salt analogues (0.63 and 0.57 nmol, respectively) compared to Mg(ClO₄)₂ alone (6.27 nmol). It should be noted that no H₂O₂ was detected in the ACN extract of the radiolyzed salt analogue that did not contain perchlorate. This suggests that, if metal peroxides and/or superoxides originate from oxygen that may be released by the radiolysis of ClO₄ ⁻ or SO₄ ²⁻ in the analogue, they are not associated with the analogue ions Na⁺, Ca²⁺, and Mg²⁺ because MgO₂, CaO₂, and Na₂O₂ were shown to be partly soluble in ACN (Table 2). The CO₃ ²⁻ component of the complete salt analogue may not play a role in the generation of metal peroxides and superoxides, since radiolysis of CO₃ ²⁻ has been associated with the formation of the carbonate radical (^•CO₃ ⁻) after CO₃ ²⁻ (or HCO₃ ⁻) reaction with ^•OH radicals (Haygarth et al., 2010). On the other hand, H₂O₂ detected in the ACN extract of radiolyzed Mg(ClO₄)₂ could have resulted from the hydrolysis of MgO₂ due to the fact that its concentration also decreases upon heating (Table 2).

H₂O₂ cannot be generated directly from irradiated Mg(ClO₄)₂ simply because it lacks protons, although, if present, trace water could be a source of protons. However, H₂O₂ is produced during aqueous radiolysis of perchlorate (Konstantatos and Katakis, 1967), most likely from the dismutation of free O₂ ^•− and not through the hydrolysis of metal peroxides, because they (e.g., MgO₂) do not form under aqueous conditions without preexistence of H₂O₂. H₂O₂ produced by radiolyzed Mg(ClO₄)₂ and the complete salt analogue may not have derived solely from the radiolysis products ClO₂ ⁻ and ClO⁻, because controls NaClO₂ and NaClO produce ∼80- and ∼150-fold higher concentration levels of ACN_alk-CE-extracted H₂O₂, respectively, than unheated radiolyzed Mg(ClO₄)₂ (Table 3). On the other hand, the high levels of H₂O₂ produced by control NaClO as compared to those produced by radiolyzed Mg(ClO₄)₂ suggest that the minor fraction of the initial ClO⁻ that survived heating in radiolyzed Mg(ClO₄)₂ may explain the low levels of H₂O₂. Assuming that radiolyzed Mg(ClO₄)₂ produces ClO⁻ at same concentration as Ca(ClO₄)₂, which is a reasonable assumption given they have the same G° value for ClO⁻ (0.03) (Prince and Johnson, 1965; Quinn et al., 2013), the ClO⁻ in radiolyzed Mg(ClO₄)₂ that survived heating corresponds to 0.43 nmol. Based on the 1.7 mol % H₂O₂ produced by NaClO (Table 3), 0.43 nmol of ClO⁻ would be expected to produce 0.007 nmol H₂O₂, which is 14-fold lower than the experimentally measured value (0.1 nmol; Table 1). This difference in H₂O₂ production by radiolyzed Mg(ClO₄)₂ may be attributed to contribution by ClO₂ ⁻, which is thermally stable. Given that the Mg(ClO₄)₂ G° value for ClO₂ ⁻ production is 0.14 (Prince and Johnson, 1965), this would correspond to the production of ∼80 nmol ClO₂ ⁻ in the analogue; and considering the production of 9 nmol H₂O₂ per 1 μmol NaClO₂ (Table 3), the assumed 80 nmol ClO₂ ⁻ in heated radiolyzed Mg(ClO₄)₂ will produce ∼0.7 nmol H₂O₂, which is 7-fold higher than the experimentally measured value (0.1 nmol; Table 1).

Considering that 6.27 nmol H₂O₂ was produced by 6.5 mg radiolyzed Mg(ClO₄)₂, the actual 82 mg sample that was dissolved in 1.6 mL ACN_alk-CE to determine experimentally this value corresponds to 50 μ M H₂O₂. On the other hand, the maximum concentration of commercial MgO₂ dissolved in ACN_alk-CE was 5.6 μ M (data not shown); thus, the same concentration of H₂O₂ would have been produced from its hydrolysis reaction [MgO₂ + 2H₂O → Mg(OH)₂ + H₂O₂]. The ∼9-fold higher concentration of H₂O₂ being produced by radiolyzed Mg(ClO₄)₂ may suggest the following: (i) either the radiolysis-generated MgO₂ is more readily dissolved if present in the crystal lattice of radiolyzed Mg(ClO₄)₂ (which was found to be readily dissolved in ACN) or (ii) H₂O₂ comes from a Mg-superoxide form such as Mg(O₂)₂ [its crystal structure has been identified (Bakulina et al., 1970)], which may also be equally ACN soluble, or (iii) both.

It has been suggested that radiolyzed ClO₄ ⁻ decomposes mainly to chlorate ions (ClO₃ ⁻) and O atoms either as aqueous solution (Katakis and Allen, 1964) or in solid form (Prince and Johnson, 1965). During the radiolysis of ClO₄ ⁻, production of ClO₃ ⁻ and ClO₂ ⁻ appears to involve oxidation of hypochlorous acid (HOCl) or chlorous acid (HClO₂) via direct electron transfer, followed by reaction of the resulting chlorine monoxide (ClO^•) or chloroperoxy radical (ClOO^•) with ^•OH (Hubler et al., 2014). Moreover, ClO₂ ⁻ acts as redox couple with the other radiolysis product ClO₂ ^•, via 1 e⁻ reduction (Stanbury and Lednicky, 1984) by, for example, ^•OH [ClO₂ ⁻ + ^•OH → ClO₂ ^• + OH⁻ (Klaning et al., 1985)] or O₂ ^•− [ClO₂ ^• + O₂ ^•− → ClO₂ ⁻ + O₂ (Huie and Neta, 1986)].

The source of residual O produced by radiolyzed perchlorate has been thought to be initially associated with ClO₂ in the ClO₄ ⁻ ion and is not all available as gaseous O₂. Considerations of electrical neutrality would favor the simultaneous formation of a metal superoxide [thus O₂ ^•−, being in the form of, e.g., Mg(O₂)₂ (Bakulina et al., 1970)], peroxide (e.g., as MgO₂), or oxide (e.g., as MgO) by the oxygen of ClO₂ (Prince and Johnson, 1965). HO₂ ^• and H₂O₂ generated by O atoms detached from radiolyzed aqueous ClO₄ ⁻ (resulting in the formation of ClO₃ ⁻), presumably reduced via H attack, have been indirectly identified (using ions of iron, cerium, and thallium as scavengers) (Katakis and Allen, 1964; Katakis and Konstantatos, 1968). Another route for O₂ ^•− formation may be initiated by photoisomerization of ClO₂ ^• and its conversion to chloroperoxyl radicals (ClOO^•) (Arkell and Schwager, 1967; Raghunathan and Sur, 1984). These upon reaction with Cl^• can be converted to chlorine monoxide radical (ClO^•, via the reaction ClO₂ ^• + Cl^• → 2ClO^•) (Enami et al., 2006), which, in turn, can react with ^•OH and generate protonated superoxide radicals via the reaction ClO^• + ^•OH → HO₂ ^• + Cl^• (Chang et al., 2004), and recycle Cl^• for further production of O₂ ^•−. Moreover, H₂O₂ can react with ^•OH and generate protonated superoxide radicals via the reaction H₂O₂ + ^•OH → HO₂ ^• + H₂O (Christensen et al., 1982; Wang et al., 2012). Other indirect mechanisms for H₂O₂ production could involve ^•OH and O₂ ^•−, both shown to be produced by radiolyzed ClO₄ ⁻ (Table 1), by ^•OH fusion [^•OH + ^•OH → H₂O₂ (Lutze et al., 2015)], and by dismutation of O₂ ^•−, protonated or not [2HO₂• → 2H₂O₂ + O₂ (Kurylo et al., 1986) or 2O₂ ^•− + 2H₂O → 2OH⁻ + H₂O₂ + O₂ (Georgiou et al., 2015)].

Given that H₂O₂ was produced by both controls NaClO₂ and NaClO (Table 3), there could also be indirect sources of H₂O₂ that may produce it upon reaction with the various Cl-based products of perchlorate radiolysis such as ClO₂ ⁻ and ClO⁻, free or both complexed with Mg [Mg(ClO₂)₂·6H₂O (Villars et al., 2011) and MgClO (since thermally stable Mg⁺ compounds are possible (Green et al., 2007)), respectively].

Another possible source of O could have been either the radiolyzed CO₂ (CO₂ → CO + O) (Lind and Bardwell, 1925; Harteck and Dondes, 1955; Watanabe et al., 2007) that is present in the glass ampule of the radiolyzed samples, or the SiO₂ of the glass ampule itself. However, neither of these sources produced O that could react with the radiolyzed incomplete salt analogue and produce traceable ROS, because this salt analogue did not generate O₂ ^•− or H₂O₂ (Table 1). Moreover, the positive holes created in silica by γ radiation are irreversibly neutralized by H₂ (giving its characteristic color as they absorb at 320 and 530 nm) (Ogura et al., 1974).

4.3. ^•OH generation by radiolyzed Mg(ClO₄)₂ and Phoenix site salt analogue

Another important finding of the present study is that the Phoenix site salt analogue acts as a source of ^•OH only upon γ radiolysis. Interestingly, ^•OH is produced at quite high and near equal levels in the absence and presence of ClO₄ ⁻, and it is still produced upon heating in both cases, although decreased by an average 2.5-fold (Table 1). On the other hand, the 6.9-fold increase in ^•OH production by heated radiolyzed Mg(ClO₄)₂ compared to an unheated one suggests the generation in radiolyzed Mg(ClO₄)₂ of certain products that may act as scavengers of ^•OH. These products may be destroyed/modified and neutralized from reacting with ^•OH (as they may do in the unheated sample, thereby decreasing its generated ^•OH levels). Considering the possible sources of ^•OH produced by the complete and the incomplete salt analogue, its ∼150-fold-high levels when compared to the 1 nmol ^•OH produced by radiolyzed Mg(ClO₄)₂ suggest a minor contribution (∼0.7%) by the ClO₄ ⁻ portion in the complete salt analogue in the production of ^•OH. It is proposed that the production of almost all ^•OH by the salt analogue (±ClO₄ ⁻) is likely to be associated with the possible radiolysis of its SO₄ ²⁻ component. A possible mechanism may involve the production of sulfites (SO₃ ²; together with elemental sulfur and sulfides), which can derive from SO₄ ²⁻ either in the presence of water ice [e.g., in Europa (Carlson et al., 2002)] or being in solid form [e.g., Li₂SO₄ (Sasaki et al., 1978)]. Sulfites could then generate ^•OH (i) via oxidation by H₂O₂ (Shi, 1994; Shi et al., 1994) detected in the present study [and also present on Mars (Encrenaz et al., 2012)] or (ii) by autoxidation (together with sulfur trioxide anion radical, SO₃ ^•−, generation) (Huie, 1986) in the presence of O₂ and H₂O [both also present on Mars, the first being possibly produced by radiolysis of perchlorate (Prince and Johnson, 1965; Quinn et al., 2013)]. Given the presence of SO₄ ²⁻ at 134 nmol in 80 mg complete salt analogue (Quinn et al., 2011), this concentration is comparable to the 149 nmol ^•OH it generates upon radiolysis (Table 1). The other prone to radiolysis salt analogue component CO₃ ²⁻ is an unlikely source of ^•OH because they react with each other to generate ^•CO₃ ⁻ radicals (Haygarth et al., 2010).

On the other hand, the minor contribution (0.7%) of ^•OH due to the presence of ClO₄ ⁻ in the complete salt analogue could have originated from the reaction of O₂ ^•− with H₂O₂ [O₂ ^•− + H₂O₂ → ^•OH + OH⁻ + O₂ (Georgiou et al., 2015)], both shown to be products of the radiolyzed complete salt analogue. This reaction may contribute to the release of O₂ by the H₂O-wetted radiolyzed perchlorate (Prince and Johnson, 1965; Quinn et al., 2013). Although O₂ ^•− was not detected to be produced by radiolyzed Mg(ClO₄)₂ (but H₂O₂ was detected), its generation cannot be excluded because, as already mentioned, the generation of metal superoxides from γ-irradiated alkali and alkaline Earth perchlorate [and Mg(ClO₄)₂] has been predicted (Prince and Johnson, 1965). Therefore, these metal peroxidants upon H₂O wetting (thus dismutation) can become the source of the detected H₂O₂ (Table 1). Production of ^•OH by the reaction between ClO⁻ and H₂O₂ at alkaline conditions (ClO⁻ + H₂O₂ → ClO^• + ^•OH + OH⁻) has also been suggested (Castagna et al., 2008), but it does not seem to be the main mechanism because we were unable to demonstrate this using NaClO and H₂O₂.

Given the oxidizing effect of NaClO₂ and NaClO on 2HTPA (Fig. 2), the indirect quantifier (and trap) of ^•OH, the question arises whether ClO₂ ⁻ and ClO⁻ [expected radiolysis products of ClO₄ ⁻ (Prince and Johnson, 1965)] will affect the quantification of ^•OH produced by the radiolyzed Mg(ClO₄)₂ and the complete salt analogue. Assuming that the radiolysis of 58.3 μmol ClO₄ ⁻ [as Mg(ClO₄)₂] generates ClO₂ ⁻, ClO⁻, and ClO₂ ^• at mol % concentration levels near equal to their G° values [0.14, 0.03, and 0.07, respectively (Prince and Johnson, 1965)], these would correspond to approximately 80, 17, and 40 nmol, respectively, per 6.5 mg radiolyzed Mg(ClO₄)₂. Since ^•OH production by the radiolyzed Mg(ClO₄)₂ alone and by the radiolyzed complete salt analogue was determined by the present study in these samples at a 1–10 mg quantity range, their corresponding concentration in ClO₂ ⁻ and ClO⁻ would not have oxidized 2HTPA substantially (Fig. 2) enough to have caused an underestimation of the actual concentration of ^•OH. The oxidizing effect of ClO₂ ^• (the other presumed radiolysis product of ClO₄ ⁻) on 2HTPA cannot be quantitatively assessed in the radiolyzed samples by commercial control analogues. However, it can be assumed to be minor due to the low (0.07) G° value of ClO₂ ^•.

4.4. Photolyzed ClO⁻ and ClO₂ ⁻ as potential sources of ^•OH, and its main generation mechanisms

Having established that ^•OH is generated at high levels by radiolyzed Mars Phoenix site salt analogue (with and without perchlorate), the present study investigated a possible photogeneration mechanism for ^•OH production by the ClO₄ ⁻ radiolysis products ClO⁻ and ClO₂ under simulated martian soil UV exposure conditions. Generation of ^•OH was studied on NaClO and NaClO₂ by exposing them to low-UV-energy (thus time-controlled) photons (using UVA at flux density of 20 μW cm⁻² up to 0.75 min exposure; Fig. 4), after having uncovered their oxychlorine products by UVC exposure at 1575 W m⁻² up to 5 min exposure (Fig. 3). These doses are ∼200-fold lower and 120-fold higher than the corresponding ones (41.5 and 13.2 W m⁻²) that reach the surface of Mars at zenith angle 0° (Cockell et al., 2000). Both NaClO and NaClO₂ produced ^•OH at rates dependent on their concentration and UVA-photolysis exposure time period (Fig. 4). Production of ^•OH by NaClO₂ is attributed to photolysis of UV-generated ClO⁻ (Fig. 3).

In hypothesizing the possible mechanisms of ^•OH generation by ClO₄ ⁻-oxychlorine radiolysis products, the accuracy of measured ^•OH was evaluated in relation to possible interfering factors. One factor is the possible oxidative effect of NaClO₂ and NaClO on 2HTPA fluorescence by which ^•OH is indirectly quantified. It was concluded that this effect is insignificant at the low concentration levels of NaClO₂ and NaClO and short UVA exposure times studied (Fig. 2). Another possible factor affecting the actual production of ^•OH by UV-photolyzed NaClO₂ could be the direct involvement of ^•OH in the production of Cl₃ ⁻ (a UVC photolysis product of NaClO₂; Fig. 3) shown in Table 4. Another important factor is the accuracy of the 2HTPA-based ^•OH standard curve used by the assay.

The involvement of the UV-photolyzed radiolysis products of perchlorate [at decreasing concentration sequence ClO₃ ⁻, O₂, Cl⁻, ClO₂ ⁻, ClO⁻, and ClO₂ ^• (Prince and Johnson, 1965)] in the production of ^•OH requires an understanding of the photolysis and interconversion reactions involved, some of which release ^•OH and others O₂ (Table 4). However, the most probable photolysis-induced ^•OH production mechanism involves HOCl or ClO⁻ (Reactions 1a or 1b, the latter via the formation of an atomic oxygen radical anion, O^•−, intermediate) (Wang et al., 2012), and/or ClO₃ ⁻ (Reaction 2) (Osiewała et al., 2013), as shown by the following corresponding reaction pathways: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \hskip+-98pt{ \rm{1a}}{ \rm{. \ HOCl }} + h {\upnu} \,{ \rm{ }}{ \to \,^ \bullet }{ \rm{OH }}{ + ^ \bullet }{ \rm{Cl}} \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \hskip+-35pt1{ \rm{b}}. \ { \rm{ Cl}}{{ \rm{O}}^ - } + { \rm{ }}{{ \rm{H}}_2}{ \rm{O }} + h {\upnu}\, { \rm{ }}{ \to \,^ \bullet }{ \rm{OH }}{ + ^ \bullet }{ \rm{Cl }} + { \rm{ O}}{{ \rm{H}}^ - } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}& ( 1{ \rm{b}} \;{ \rm{is}} \;{ \rm{the}} \;{ \rm{sum}} \;{ \rm{of}} \;{ \rm{reactions :}} \;{ \rm{Cl}}{{ \rm{O}}^ - } + h{\upnu} \,{ \rm{ }}{ \to\, ^ \bullet }{ \rm{Cl }} + {{ \rm{O}}^{ \bullet - }} \\ & \quad{ \rm{and}} \;{{ \rm{O}}^{ \bullet - }} + { \rm{ }}{{ \rm{H}}_2}{ \rm{O }}\,{ \to \,^ \bullet }{ \rm{OH }} + { \rm{ O}}{{ \rm{H}}^ - } ) \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}& 2. {\rm {ClO}_3}^ - + {{ \rm{H}}_2}{ \rm{O}} + h {\nu} \longrightarrow {^\bullet}{\rm{OH}} + { \rm{ ClO}_2}^{\bullet} + { \rm{O}}{{ \rm{H}}^- } \\ & \quad\quad (\rm Osiewa\l \ et \ al ., 2013) \end{align*} \end{document}

The stoichiometry of the above reaction pathways predicts generation of 1 mol ^•OH by 1 mol photolyzed ClO⁻. However, our experiments with photolyzed NaClO showed that 7.6 and 23 nmol produced ^•OH concentration plateaus at 16 and 45 nmol, respectively, which corresponds to a ClO⁻/^•OH 1/2 molar stoichiometry (Fig. 4a). This finding can be explained by involving the reactions of the other photolysis product ^•Cl (Cl atom) with H₂O or OH⁻ (Reaction 3a or 3b) (Lutze et al., 2015), and HO₂ ^• (Reaction 4) (Chang et al., 2004), which will produce the second mole of ^•OH required for the ClO⁻/^•OH molar ratio 1/2: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \hskip+-44pt3{ \rm{a}}.\ { \rm{ With}} \;{{ \rm{H}}_2}{ \rm{O}} :{ \rm{ C}}{{ \rm{l}}^ \bullet } + {{ \rm{H}}_2}{ \rm{O}}\,{ \to\, ^ \bullet }{ \rm{OH }} + { \rm{ C}}{{ \rm{l}}^ - } + {{ \rm{H}}^ + } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*}&( 3{\rm a \ is \ the \ sum \ of \ reactions: Cl}^{\bullet} +{{\rm{H}}_2} {\rm{O}} \longrightarrow {\rm{HClO}}^{\bullet-} +{\rm{H}}^+ \\ & \quad {\rm and} \ {\rm{HClO}}^{{\bullet-}} \rightleftarrows {^\bullet} {\rm{OH }} + {\rm Cl}^- ) \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \hskip+-62pt3{ \rm{b}}. \;{ \rm{With}} \;{ \rm{O}}{{ \rm{H}}^ - } :{ \rm{ C}}{{ \rm{l}}^ \bullet } + { \rm{ O}}{{ \rm{H}}^ - }\,{ \to\, ^ \bullet }{ \rm{OH }} + { \rm{ C}}{{ \rm{l}}^ - } \end{align*} \end{document} \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} \hskip+-65pt\;4. \;{ \rm{With}} \;{ \rm{H}}{{ \rm{O}}_2}^ \bullet :{ \rm{C}}{{ \rm{l}}^ \bullet } + { \rm{ H}}{{ \rm{O}}_2}^ \bullet \to { \rm{Cl}}{{ \rm{O}}^ \bullet }{ + ^ \bullet }{ \rm{OH}} \end{align*} \end{document}

In turn, Cl^• and ^•OH may be involved in reactions with the radiolysis products of perchlorate (Table 4).

4.5. Soil ROS on Mars

The present study suggests that martian topsoil is a continuous source of ROS generated by cosmic radiolysis and by a parallel UVA/C photolysis of ClO₂ ⁻ and ClO⁻, both products of the cosmic radiolysis of ClO₄ ⁻. The ClO₄ ⁻ radiolysis-H₂O-wetting-induced ROS generation mechanism (Fig. 5) originates from a Mars soil salt analogue (Phoenix site). This phenomenon may take place in parallel with another independent generation mechanism of soil ROS (O₂ ^•−, H₂O₂, and secondarily ^•OH), by UV-induced electron ejection from soil mineral oxides and its subsequent trapping by O₂ as O₂ ^•− (Georgiou et al., 2015), which is either released directly by radiolyzed ClO₄ ⁻ (Quinn et al., 2013) or produced indirectly by the dismutation of O₂ ^•− that is also generated by radiolyzed ClO₄ ⁻ (Fig. 5).

OPEN IN VIEWER

FIG. 5.

Mechanisms of ROS generation by perchlorate and Mars Phoenix site salt analogue exposed to cosmic/UV radiation. Cosmic (γ) radiation decomposes perchlorate to ClO₃ ⁻, O₂, Cl⁻, ClO₂ ⁻, ClO⁻, and ClO₂ ^• (in a G° value decreasing sequence) (Prince and Johnson, 1965), shown by solid arrows. Part of released residual O _x , associated with ClO₂ in the ClO₄ ⁻, is matrix-trapped as gaseous O₂, while the other part is converted to O₂ ^•− (shown to be generated by the radiolyzed Mars Phoenix site complete salt analogue; Table 1), which in the presence of martian salt metal ions (Me) would favor the simultaneous formation of metal superoxides and peroxides (Katakis and Allen, 1964; Prince and Johnson, 1965; Katakis and Konstantatos, 1968). O₂ ^•− (free and present in metal superoxides) could dismutate to O₂ (an additional source of oxygen) and H₂O₂ (that could be also released by the hydrolysis of metal peroxides), which could then react with each other to generate ^•OH (Georgiou et al., 2015). Cosmic radiation could also radiolyze the salt analog component SO₄ ²⁻ to the intermediate sulfite, SO₃ ²⁻, which upon oxidation by H₂O₂ (Shi, 1994; Shi et al., 1994) or autoxidation (Huie, 1986) in the presence of H₂O and O₂ (released, e.g., by radiolyzed ClO₄ ⁻) will generate ^•OH (as shown with the incomplete salt analogue, Table 1). Simultaneous UV photolysis of the ClO₄ ⁻ radiolysis product ClO₂ ⁻ produces additional ClO⁻ and ClO₂ ^•, and also Cl₃ ⁻ (Fig. 3) as shown with dotted arrows. Then, upon UV photolysis, ClO⁻ (and ClO₂ ⁻ via ClO⁻; Fig. 4) as well as ClO₃ ⁻ in the presence of H₂O will generate ^•OH (ClO⁻ via the intermediate atomic oxygen radical anion, O^•−), 2 mol of which can combine to produce 1 mol of H₂O₂.

The production of ^•OH and H₂O₂ by radiolyzed ClO₄ ⁻ may contribute to the oxidative alteration of organics present in martian surface material. CO₂ can be produced by the reaction of organics with either H₂O₂ or ^•OH. H₂O₂ can decompose carbohydrates to formic acid and ultimately to CO₂ (Payne and Foster, 1945), while ^•OH can produce CO₂ through various oxidation reactions with organics [e.g., with methanol (Gehringer et al., 1988; Munter, 2001) and with the products of oxidized organics carboxylic acids such as fumaric acid (Shi et al., 2012)].

Considering the seasonality of the hydrated perchlorate-associated recurring slope lineae phenomenon (Ojha et al., 2015) and the consequent periodic hydration of the ROS generating γ-radiolyzed/UV-photolyzed oxychlorine compounds and UV-radiated soil mineral oxides over short timescales, this raises the question of the potential lifetime of these ROS when wet. In such a case, ROS lifetime may be short and may not be an issue for habitability. On the other hand, surface water flows would also be exposed to UV when wet, which should generate ROS, perhaps indicating poor preservation potential. Analyses of the martian surface performed with the SAM instrument on the Mars Science Laboratory have established the presence of significant quantities of organic matter (Ming et al., 2014; Freissinet et al., 2015). However, the chemical identities of the parent organic compounds present in the tested samples are obscured due to a thermal decomposition process induced by the presence of perchlorate (Ming et al., 2014; Freissinet et al., 2015). This points to the need for consideration of sample analysis methods in future missions. Although levels of ROS are expected to be low in comparison to perchlorate, sample cleanup processes (e.g., solid phase extraction) or non-aqueous chemical derivatization may be required in the future prior to implementation of thermal or aqueous-based analysis methods. Additionally, analytical methods planned as part of the Mars Organic Molecule Analyzer (MOMA) on the upcoming ESA ExoMars mission include laser desorption mass spectroscopy (LDMS), which has been shown to be capable of the detection of trace organics in perchlorate-containing Mars analogues (Li et al., 2015) and likewise should be compatible with the detection of organics in the presence of ROS-containing martian soils. The other instrument suitable for detection of organics, planned for the NASA Mars 2020 mission, is the Scanning Habitable Environments with Raman and Luminescence for Organics (SHERLOC) (Beegle et al., 2014). SHERLOC uses a deep UV resonance Raman and fluorescence spectrometer utilizing a 248.6 nm deep UV laser for noncontact sample analysis. The technique may modify organics by inducing reactions with oxychlorine species (Beegle et al., 2014). In light of the present study, modification of organics by ROS generated by UV irradiation of oxychlorine products may not be excluded. Moreover, UV bleaching may also be of concern.

The present study could partly explain the finding by the Curiosity rover of high Mn abundances (>25 wt % MnO) in fracture-filling materials that crosscut sandstones in the Kimberley region of Gale Crater of Mars (Lanza et al., 2016), because their proposed accumulation as Mn-oxide deposits on Earth environments requires water and highly oxidizing conditions. The latter can be provided on Mars by the radiation-driven formation of ROS in the oxychlorine compounds of topsoil. Another possible source of ROS could be provided by the soil superoxides/peroxides that are possibly generated on Mars by the UV-irradiated topsoil, as they can be generated on Earth by UV-exposed arid desert soils (Georgiou et al., 2015). The latter alternative ROS source and the fact that not all radiated oxychlorine products could produce ROS (Fig. 5) can explain the reported lack of correlation between Mn and the element Cl (Lanza et al., 2016).

The impacts and distributions of ROS on the surface of Mars can be addressed in more specific terms by field instrumentation for the in situ detection and quantification of the metal superoxides (O₂ ^•−) and peroxides (H₂O₂) present in martian topsoil. A methodology for the development of such an instrument is already available and is based on the enzymic/non-enzymic conversion of O₂ ^•− and H₂O₂ to stoichiometric O₂ (g) and, thus, their quantification by an O₂ electrode (Georgiou et al., 2016).

To more faithfully recreate the cosmic ray particle bombardment onto the martian surface and near subsurface, similar irradiation experiments are planned using electron and proton bombardment combined with X-ray photoelectron spectroscopy analysis. For a more accurate simulation of martian soil, the Phoenix site salt analogue could be mixed with iron oxides, other oxides, and other expected soil components (e.g., clays), which would likely affect the production (and chemistry) of the oxychlorine radiation-generated ROS. This is supported by the ∼6-fold decreased rate of ^•OH generation by H₂O extracts of desert soil compared to its absence (Georgiou et al., 2015), possibly due to the scavenging action of metal (Me) ions [^•OH + Me ⁿ⁺H₂O → Me ⁿ⁺OH → Me⁽ⁿ⁺¹⁾⁺ + OH⁻ (Buxton et al., 1988)].

4.6. Relationships to the Viking LR and GEx experiments

Our results show that radiolysis of Mg(ClO₄)₂ in a 7 torr CO₂ atmosphere produces peroxide, but not superoxide. ^•OH is also produced presumably due to trace amounts of water associated with the Mg(ClO₄)₂ sample. Heat treatment of the sample prior to analysis resulted in a >98% decrease in peroxide concentration and a corresponding increase (number of moles) in ^•OH. In the Viking GEx experiments, the production of 70–700 nmol of O₂ (g) was measured when martian surface samples were humidified or wetted (Oyama and Berdahl, 1977). The total quantity of ROS sum (O₂ ²⁻, O₂ ^•−, and ^•OH) measured in the irradiated (500 kGy) Mg(ClO₄)₂ (without Phoenix salt analogue) in this study totaled ∼7 nmol. However, the 1 nmol of ^•OH measured is not an accurate representation of the ^•OH content. The reason for this is that ^•OH is indirectly quantified (via scavenging by TPA). Therefore, we estimate (data not shown) that the level of ^•OH in the sample is ∼30 nmol and that the total quantity of ROS in the sample is closer to the lower end of the range of O₂ (g) levels observed in the Viking GEx experiment (70–700 nmol), assuming similar levels of perchlorate at both the Viking and Phoenix sites. Importantly, the amount of ROS (primarily ^•OH) measured in the Phoenix salt (without perchlorate) analogue was within the range of O₂ (g) levels. Along with the observed thermal stability, this points to the likely extensive formation of ROS on the surface of Mars across diverse mineral matrices, which suggests that there may be additional sources of oxygen [e.g., SO₄ ²⁻, soil mineral oxides (Georgiou et al., 2015)] that may have contributed to the GEx O₂ (g) releases. It should be noted, though, that the kinetics of O₂ (g) formation from ^•OH humidification or wetting was not examined in the present study. Additionally, the amount of ROS measured in our experiments represents only a small fraction of the total free oxygen produced by perchlorate radiolysis. The ROS mol % fraction measured relative to total Mg(ClO₄)₂ in the current study was 0.01. In contrast, Quinn et al. (2013) measured a 1 mol % fraction of O₂ (g) released from irradiated (500 kGy) Ca(ClO₄)₂ upon sample wetting. The expected total amount of free oxygen produced in the γ-irradiated Mg(ClO₄)₂ examined in this study is ∼2 μmol, based on a G° value of 2.62 (Prince and Johnson, 1965), which correlates to a measured ROS mol % fraction of 0.4 (relative to the expected total oxygen production measured as O₂). While this may imply that ultimately the majority of oxygen produced by perchlorate radiolysis is stabilized and trapped in, or released from, the perchlorate as O₂ (g), our results suggest that radiolysis pathways dictate that the formation of O₂ (g) occurs via ROS metastable intermediates, which ultimately form, and coexist, with trapped O₂ (g). The significance of this is that the ROS formed during radiolysis of martian surface materials may not only serve as the soil O₂ precursor that ultimately was measured as O₂ (g) in the Viking GEx experiment, but also, due to their high reactivity, likely act to chemically alter both the mineral and organic fractions of martian surface materials in situ.

In the Viking LR experiments, when 0.5 cc of surface soil material was wetted with an aqueous nutrient media that contained ¹⁴C-labeled organics, ∼30 nmol of ¹⁴C-labeled gas (presumably ¹⁴CO₂) was released into the LR test cell headspace (Levin and Straat, 1977). Table 1 shows that the levels of H₂O₂ and O₂ ^•− produced in our experiments were lower than the ∼30 nmol ¹⁴C-release observed in the Viking LR, but more importantly these samples were not tested to demonstrate that they reproduced the kinetics of the LR ¹⁴C-gas release. Production of ^•OH in the irradiated Phoenix salt analogues (both with or without perchlorate) can be discounted as a possible explanation of the LR experiment due to its thermal stability. However, assuming that production of low levels of H₂O₂ by heated radiolyzed Ca(ClO₄)₂ is similar to those produced by heated radiolyzed Mg(ClO4)₂ (61-fold decreased H₂O₂ production of heated over unheated; Table 1), these very low H₂O₂ levels (and possibly low levels of ^•OH) may contribute to CO₂ production, although not at levels seen in the Ca(ClO₄)₂ radiolysis study (Quinn et al., 2013) that resulted from the formation and spontaneous decomposition of chloroaniline.

An important finding of the current study is the reduced recovery (55.6%) of ClO⁻ after heating NaClO (water removal at 20–50°C) and a 97% reduction in ClO⁻ recovery after heating NaClO for 3 h at 160°C). Levin and Straat (2016) pointed out that Quinn et al. (2013) did not attempt to demonstrate that hypochlorite matches the observed thermal stability profile of the LR oxidant. In the Viking LR experiment, the magnitude of the ¹⁴CO₂ release was reduced by ∼70% when heated to 46°C for 3 h and was nearly completely eliminated when heated to 51°C for 3 h. Although the current results do not offer a perfect match to the Viking LR, these experiments were not designed to be a rigorous test of LR results. The results do show, however, that the thermal stability profile of NaClO is in the range of what was observed in the LR experiment. Further experiments are planned to characterize the thermal stability of hypochlorite and other oxychlorine species in the context of the LR experiments.