Capacity of Chlorate to Oxidize Ferrous Iron: Implications for Iron Oxide Formation on Mars

: Chlorate is an important Cl-bearing species and a strong potential Fe(II) oxidant on Mars. Since the amount of oxychlorine species (perchlorate and chlorate) detected on Mars is limited ( < ~1 wt.%), the e ﬀ ectiveness of chlorate to produce iron oxides depends heavily on its oxidizing capacity. Decomposition of chlorate or intermediates produced during its reduction, before reaction with Fe(II) would decrease its e ﬀ ective capacity as an oxidant. We thus evaluated the capacity of chlorate to produce Fe(III) minerals in Mars-relevant ﬂuids, via oxidation of dissolved Fe(II). Each chlorate ion can oxidize 6 Fe(II) ions under all conditions investigated. Mass balance demonstrated that 1 wt.% chlorate (as ClO 3 − ) could produce approximately 6 to 12 wt.% Fe(III) or mixed valent mineral products, with the amount varying with the formula of the precipitating phase. The mineral products are primarily determined by the ﬂuid type (chloride- or sulfate-rich), the solution pH, and the rate of Fe(II) oxidation. The pH at the time of initial mineral nucleation and the amount of residual dissolved Fe(II) in the system exert important additional controls on the ﬁnal mineralogy. Subsequent diagenetic transformation of these phases would yield 5.7 wt.% hematite per wt.% of chlorate reacted, providing a quantitative constraint on the capacity of chlorate to generate iron oxides on Mars.

In addition, ClO 3 − reduction to chloride by dissolved Fe(II) occurs sequentially through six stepwise reactions [69], generating a series of reactive, intermediate, chlorine-bearing species. The relative rates of decomposition and oxidization of Fe(II) by these species is unclear. If chlorate or its intermediate reduction products partially break down before they fully react with Fe(II), then the overall capacity for iron oxidation by chlorate is lower than the theoretical 6:1 molar ratio. While this excess chlorate consumption in our prior study might have been an analytical error [49], if valid, then the amount of Fe(II) that ClO 3 − can oxidize would be halved.
Since the amount of ClO 3 − expected to be present in Martian near-surface systems is limited (<1 wt.%, not accounting for regeneration processes [70][71][72][73][74][75]), understanding its actual capacity to oxidize dissolved Fe(II) in Mars-relevant fluids might better constrain iron oxidation processes and the resulting products. The number of moles of Fe(II) oxidized per mole of ClO 3 − ions relative to its maximum, theoretical capacity of 6:1 (Equation (1)), is referred to as the "stoichiometric efficiency". The rate of Fe(II) oxidation by chlorate is faster than by O 2 or through UV-photooxidation, by orders of magnitude in fluids similar to the Martian surface waters [49]. However, the amount of ClO 3 − in Martian soil or sediments could be an important limiting factor in determining its relative contribution as an Fe(II) oxidant on Mars. In addition, this stoichiometric efficiency potentially varies with fluid composition (e.g., chloride concentration, pH) and will be affected by the aqueous speciation of chlorine and dissolved Fe(II  [49], which explicitly assumed full chlorate efficiency, to compare the practical oxidation capacity of chlorate against its theoretical maximum. Our results would allow us to examine the applicability, as well as the limits of the kinetic model, in conditions that are geochemically different than those in which it was originally parameterized. The resulting mineral products were identified and quantified to determine the different iron oxides produced by chlorate as a function of fluid composition.
The reactors to study the reaction kinetics contained 40 mL solutions of~10 mmol L −1 Fe(II), with either 1.  ratios that are not exactly equal to the intended ratios. Mars-relevant background salts magnesium chloride or magnesium sulfate [57,[76][77][78][79][80][81] were used to serve as ionic strength buffers (~100 mmol L −1 ) and provide anions with a different ability to complex dissolved Fe(II), affecting the reaction rates and mineral products. Owing to the similar reaction rate in chloride-and perchlorate-rich fluids [49], experiments in magnesium perchlorate fluids were not investigated to avoid redundancy. The initial pH values were set to 7, 5, or 3 using 1 mol L −1 hydrochloric acid and sodium hydroxide. These initial pH values were chosen to represent a range of near-neutral to acidic pH values that were thought to represent most of the fluid pH conditions on Mars [82][83][84]. The acidity on Mars was controlled primarily by Fe(II) oxidation and Fe(III) mineral precipitation [46], and therefore the experimental pH was allowed to drift freely as a response to Fe(II) oxidation and Fe(III) precipitation. The pH and Fe(II) concentration were measured throughout the course of the reactions, the latter via spectrophotometry, following complexation by ferrozine [85]. For dissolved Fe(II) determination by spectrophotometry, the subsamples from the experimental solutions were extracted and centrifuged, before complexation with ferrozine, and the absorbance was corrected using a blank containing DI water and ferrozine. Chlorate concentrations were measured at the start and the end of the experiment, using a Dionex Integrion high pressure ion chromatograph (Thermo Scientific, Sunnyvale, CA, USA) equipped with an IonPac AS11 analytical column, ADRS 600 suppressor, and conductivity detector with a 12 mmol L −1 KOH eluent. The total iron [Fe(II) and Fe(III)] in the solutions after the end of the experiments were measured using an iCAP 7400 Duo inductively coupled plasma optical emission spectroscopy (Thermo Scientific, Madison, WI, USA).
A kinetic rate law model for dissolved Fe(II) oxidation by chlorate [63][64][65] was previously implemented in The Geochemist's Workbench module React [49], using a modified LLNL database [86,87]. This kinetic model was employed in the present study to simulate the reaction of chlorate with Fe(II), using the starting compositions of the experiments. Details of the kinetic model, rate constants, and the activation energy can be found in our previous paper [49]. a Sample name code: The abbreviated names of each experiment represent the fluid composition (either Cl = chloride, or S = sulfate), ratio of iron (II) to chlorate (either ≈6:1 or ≈10:1), concentration of Fe(II) in mmol L −1 , concentration of background salt (MgCl 2 or MgSO 4 ) in mmol L −1 and the initial pH. b Concentration~100 days after the start of the experiments. c Below detection limit (0.012 mmol L −1 ). d pH~100 days after the start of the experiments.

Mineral Precipitation Studies
The Fe(II) oxidation products were studied for a range of fluid compositions and Fe(II):ClO 3 − ratios by conducting experiments at larger volumes (150 mL) to facilitate greater solid phase production needed for characterization ( Table 2). The experimental solutions for the higher volume mineral precipitation experiments were prepared in serum bottles and sealed inside the anaerobic chamber and then placed on an orbital shaker outside the chambers, operating at~160 rotations per minute for~100 days. A control sample containing dissolved Fe(II) but no chlorate at pH 7 was similarly prepared in a serum-bottle reactor to verify the effectiveness of using sealed serum bottles, outside the anaerobic chamber for anaerobic experiments. The suspensions were then filtered (0.22 µm pore size MCE membrane) and dried inside the anaerobic chamber, using a vacuum desiccator. The solid samples were analyzed by powder X-ray diffraction (XRD) using a D8 Advance diffractometer (Bruker AXS, Karlsruhe, Germany) equipped with a Cu Kα source and a position sensitive, energy-dispersive LynxEye XE detector. Diffrac.Eva version 4.0 (Bruker AXS, Karlsruhe, Germany) was used for mineral identification, and the Profex version 4.2 [88] interface to BGMN [89] was used for quantitative phase analysis via the Rietveld method.    Figure 1, Table 1). The pH dropped to~2 in chloride-and~2.5 in sulfate-rich fluids, in response to Fe(II) oxidation and subsequent Fe(III) hydrolysis. Similar to Mitra and Catalano [49], the rate of Fe(II) oxidation was slower in sulfate-than chloride-rich fluids, despite the greater Fe(II) to chlorate ratios in the present study. Fe(II) oxidation in chlorate-equivalent systems for all fluids and initial pH conditions displayed about 90% of the maximum stoichiometric extent of reaction within 100 days. The chlorate-equivalent systems experienced a greater extent of Fe(II) oxidation than chlorate-deficient systems, due to the stoichiometrically greater amount of chlorate in the system. The rate of Fe(II) oxidation by chlorate was proportional to the chlorate concentration [63], explaining the slower oxidation rate in chlorate-deficient systems than the chlorate-equivalent and "chlorate-excess" systems ([Fe(II)]/[ClO 3 − ] ≈ 1) [49]. Unlike chlorate-equivalent systems that contain nearly enough chlorate to oxidize all dissolved Fe(II), the chlorate-deficient systems were expected to contain excess substantial Fe(II) ions after the consumption of all available chlorate in the solutions. The small amount of dissolved Fe(II) unoxidized in the chlorate-equivalent experiments was due to the slightly greater [Fe(II)]/[ClO 3 − ] than 6 and the insufficient time to undergo full oxidation. The pH evolution trend matched the Fe(II) reduction trend and the chlorate-deficient systems showed a slower pH decrease than the chlorate-equivalent systems. This effect was expressed more clearly in sulfatethan in chloride-fluids, owing to the slightly slower Fe(II) oxidation rate. At the end of the experiments, the chlorate in all reactors was below the detection limit (0.012 mmol L −1 ) ( Table 1).

Iron Oxidation in Chlorate-Equivalent vs. Chlorate-Deficient Systems
To further evaluate the stoichiometric efficiency of chlorate, the kinetic data were compared to the predictions of the rate model from Mitra and Catalano [49], which followed the following rate law [63][64][65] This model explicitly assumed a 100% efficiency, i.e., 1 mole of ClO 3 − oxidized 6 moles of Fe(II). In other words, the kinetic model provided the ideal extent and rate of Fe(II) oxidation, chlorate consumption, and pH evolution, as a function of time corresponding to each experiment. The results of the experimental studies were compared to the model results to understand the practical efficiency of chlorate to oxidize Fe(II) in experimental conditions. Both Fe(II) concentration and the pH of the experiments were in close agreement with predictions of our previously-parameterized kinetic model [49], across near-neutral to acidic pH conditions in both chloride-and sulfate-rich fluid systems (Figure 1). The kinetic model also predicted near-complete consumption of chlorate ( Figure S2), as observed in the experiments (Table 1) at the end of~100 days. This model-data comparison demonstrated that chlorate expressed its maximum capacity for oxidizing dissolved Fe(II) in Mars-relevant fluids.

Mineral Products in Chlorate-Equivalent vs. Chlorate-Deficient Systems
Fe(II) oxidation by chlorate and subsequent Fe(III) hydrolysis yielded diverse mineral products in the corresponding~100 day duration replicate mineral precipitation experiments ( Figure 2, Table 2). The final chlorate concentration in all replicate experiments dropped below the detection limit of the ion chromatograph ( Table 2). The final pH values of the replicate solutions matched with the final pH values of the corresponding kinetic experiments (Table 1), except the solutions starting at near-neutral solutions, in which minor deviation in the final pH was observed. The chloride fluids that started at pH 3 primarily produced dissolved Fe(III), and filtering the solution did not yield sufficient minerals for characterization. The final pH of these solutions dropped below pH 2, facilitating greater solubility of the generated Fe(III) and leading to decreased amounts of mineral precipitation. A portion of the minerals that formed in these solutions were likely nanoparticulates and passed through the filter due to their small particle size. Discoloration indicated that particles were also retained by the filter membrane but these remained embedded and could not be recovered for analysis. The remainder of the precipitate was too limited in quantity to recover and analyze. All remaining experiments yielded mineral precipitates that depended on the fluid composition, the extent of Fe(II) oxidation, and the reaction rate ( Figure 2).

Minerals Produced in Chloride-Rich Fluids
The chloride-rich fluid solutions that started at pH~7, unlike pH 3, produced a mixture of Fe(III)-and mixed-valent Fe(II/III) minerals (Figure 2a,b, Table 2, and Table S1). Chlorate-equivalent experimental solutions that started at pH 7 produced a mixture of lepidocrocite, nanocrystalline goethite, and nonstoichiometric magnetite. The proportion of lepidocrocite in the mineral precipitate was greater for higher concentrations of Fe(II) and chloride in the solution. In contrast, chlorate-deficient systems formed a mixture of green rust, goethite, and magnetite. The high-concentration chlorate-deficient system showed formation of magnetite with minor goethite. The chlorate-deficient systems maintained substantial unreacted Fe(II) and produced slower oxidation rates and relatively higher pH values, which facilitated the predominance of magnetite over other Fe(III) minerals.

Minerals Produced in Sulfate-Rich Fluids
All experiments conducted in the sulfate-rich fluid systems yielded mineral precipitates containing either magnetite, goethite, or schwertmannite (Figure 2c,d, Table 2, and Table S1). The SO 4 -rich fluids did not precipitate lepidocrocite or green rust. The experimental solutions that started at pH 3 produced either schwertmannite or goethite, depending on the Fe(II) concentration. Solutions with high solute concentration ([Fe(II)] ≈ 100 mmol L −1 ) formed schwertmannite as the sole phase. The final pH of these solutions dropped below pH~2.3 and no other Fe(III)-bearing phases co-existed with schwertmannite. The solutions containing low solute concentrations ([Fe(II)] ≈ 10 mmol L −1 ) starting at initial pH 3, precipitated nanocrystalline goethite instead of schwertmannite. Experimental solutions that started at pH 7 produced magnetite or more crystalline goethite. The proportion of crystalline goethite also increased in solutions that started at near neutral solutions.
Sulfate-rich solutions containing greater than~10 mmol L −1 Fe(II) precipitated magnetite as the predominant mineral in chlorate-deficient systems at neutral pH. On the other hand, analogous chlorate-equivalent systems did not precipitate magnetite. We investigated a chlorate-equivalent solution with even lower solute concentrations ([Fe(II)] ≈ 5 mmol L −1 ) to further decrease the rate of Fe(II) oxidation. Magnetite precipitated with minor goethite in this dilute chlorate-equivalent system, when the initial pH was 7 but did not form any detectable mineral precipitate in solution that started at pH 3. The precipitation of magnetite was favored over goethite in this system, owing to the very slow Fe(II) oxidation rate, due to low initial [Fe(II)] and the mild acidic nature of the final fluid (pH final ≈ 4.6).

Fe(II) Oxidation Capacity of Chlorate
Our results demonstrated that chlorate expresses the full theoretical capacity to oxidize Fe(II) in Mars-relevant fluid systems. The final observed dissolved Fe(II) concentration in all experimental solutions at the end of~100 days were similar to the calculated final values predicted in the same time period by the kinetic model (Tables 1 and 2), which assumed full oxidation efficiency of chlorate ions. The close agreement between the experiment and the model (Figure 1) demonstrated that each chlorate ion could oxidize six dissolved Fe(II) ions without substantial loss via side-reactions. Therefore, any chlorate present on the surface of Mars is expected to be fully available to react with Fe(II), providing a substantial oxidizing capacity. This also confirmed the applicability of the previously published rate law model [49] to accurately predict the rate of Fe(II) oxidation by chlorate in diverse systems. The model can therefore be confidently used to simulate Fe(II) oxidation by chlorate in diverse Mars-relevant fluids.

Fluid Composition
The fluid composition, specifically the dominant anion present, exerts strong control over the mineral product by influencing the mineral formation pathways [24] and by providing essential components for some minerals (e.g., chloride for akaganeite, sulfate for schwertmannite and jarosite) (Figure 3). The effect of fluid composition on the nature of mineral product formation is shown in our results ( Figure 2) and our earlier study [49]. Iron oxidation by chlorate in chloride-rich fluids produced either lepidocrocite, goethite, magnetite, akaganeite, or a mix of these minerals. Chloride and other halogenides (bromide, iodide, fluoride) promote lepidocrocite formation [90,91], and akaganeite forms in acidic chloride-rich fluids [24,[92][93][94][95]. Similarly, in sulfate-rich solutions iron oxidation by chlorate produced either goethite, magnetite, schwertmannite, jarosite, or a combination of these minerals. The precipitation of jarosite (hydronium-and natro-) was observed by Mitra and Catalano [49]. Sulfate ions also favors goethite formation over lepidocrocite [24], hence the predominance of goethite in sulfate-rich fluids. Schwertmannite forms as a precursor mineral to jarosite and goethite in Fe(II) oxidizing sulfate-rich fluids [96][97][98].

pH Evolution
The solution pH also played an important role in determining the type and proportion of the mineral precipitate (Figure 3). Highly acidic systems inhibited the precipitation of substantial Fe(III) minerals because of mineral nucleation inhibition and the weak supersaturation of iron oxides at low pH (<2) (Figure 2, Table 2). When minerals do form in substantial quantities, the final pH alone does not determine the parameter of the mineralogy, as the data clearly demonstrated that this was affected by the pH at the time of nucleation. The fluids at the end of the experiments were acidic; this suggest that phases such as akaganeite [Fe 8 O 6.4 (OH) 9.7 Cl 1.3 or β-FeO(OH,Cl)] and jarosite [XFe 3 (OH) 6 (SO 4 ) 2 , X = K + , Na + or H 3 O + ] should form. However, green rust and magnetite, which typically form in near neutral and alkaline conditions [24], formed in some cases, with goethite and lepidocrocite also occurring. This indicated that the circumneutral pH at the time of mineral nucleation had a substantial effect on the fate of Fe(III).
The relative abundance of goethite and schwertmannite in sulfate-rich systems was also primarily controlled by the solution pH. Schwertmannite is a common mineral in acid sulfate systems and was shown to occur between pH 4.7 to 1.9 [98]. It is a metastable precursor mineral forming in oxidizing sulfate-rich Fe(II) solutions in acidic systems [96,97]. The stability of schwertmannite in our experiments might thus be transient, although it is apparently stable in the studied fluids for at least 100 days.
In other studies schwertmannite was shown to transform to either goethite or jarosite, depending on the pH and sulfate concentration of the solution [97,[99][100][101][102][103][104]. In the present work, schwertmannite formed only in experimental solutions that started at pH~3 and then dropped below pH~2.3. The preservation of schwertmannite in our experiments demonstrated the important role played by the solution pH at the time of mineral nucleation, in determining the type of mineral precipitate.
The frequent occurrence of magnetite was unexpected, given the low final pH of some experiments. Notably, magnetite only precipitated in solutions that started at near-neutral pH. The solution pH at the time of magnetite nucleation was likely~4. Simulation of the exact initial composition of the experiment using the kinetic model found that the pH was at this value at the initiation of magnetite precipitation (simulation not shown). Magnetite is metastable in acidic solutions, owing to its high solubility [105,106]. It persists in our experimental solutions either due to slow dissolution kinetics [107,108], or the formation of an oxidized surface layer that inhibited further dissolution. Lattice parameters and octahedral iron occupancies from Rietveld refinements of the XRD results indicate the presence of non-stoichiometric magnetite in a number of experiments ( Figure 2). Non-stoichiometric magnetite, also referred to as "cation-deficient magnetite", had an Fe(II)/Fe(III) ratio <0.5 and a unit cell smaller than regular magnetite (8.396 Å), caused by Fe(II) oxidation to Fe(III), and the presence of vacancies in its structure [109]. The formation of non-stoichiometric magnetite might be caused either by direct oxidation of magnetite by chlorate or by the leaching of surficial Fe(II) ions of magnetite in acidic fluids (pH 2-2.5) [110]. Our experiments depicted a similar case in which magnetite at the end of the experiment was retrieved from acidic solutions.

Oxidation Rate
Geochemically similar systems can produce different mineral products due to the differences in their iron oxidation rate ( solutions investigated in our prior investigation [49] produced akaganeite, due to very rapid oxidation rates. In contrast, lepidocrocite formed in similar fluids in the chlorate-equivalent systems of the present study due to slower rate of Fe(II) oxidation ( Figure 2). Slower rates of oxidation in chloride-rich, chlorate-deficient systems induced magnetite formation over lepidocrocite for initially circumneutral pH conditions. On the contrary, during faster Fe(II) oxidation, lepidocrocite formation is preferred over magnetite ( Figure S7). The figure demonstrates how the mineral type is a strong function of the oxidation rate. A mixture of lepidocrocite and goethite is formed in near-neutral solutions if the rate of formation is slightly slower [111]. In chloride-rich fluids, systems favoring faster iron oxidation form akaganeite [49], while those yielding slower oxidation rates form lepidocrocite and magnetite.
Similar to the chloride system, for sulfate-rich fluids the chlorate-excess solutions studied previously produced favored goethite [49]. Slower oxidation rates in chlorate-equivalent and -deficient systems produced either goethite, schwertmannite, or magnetite. Relatively faster oxidation rates favored goethite (or schwertmannite, depending on initial pH) production over magnetite. The sulfate-rich fluids produced magnetite in systems with an even slower oxidation rate, which contained unoxidized Fe(II) along with oxidized Fe(III) in the solution. Magnetite formed either in chlorate-deficient systems containing low [Fe(II)] (~10 mmol L −1 ) (e.g., S-10:1-10-100-7) or in chlorate-equivalent systems with even lower [Fe(II)] (~5 mmol L −1 ) (e.g., S-6:1-5-100-7); refer to Tables 1 and 2 footnotes for explanation of the sample codes. In all systems, magnetite formed in solutions undergoing slow oxidation at near-neutral solutions ( Figure S7). Sulfate-rich solutions produced magnetite only when the rate of oxidation was orders of magnitude slower than those which formed goethite or jarosite.
Such diagenetic transformations are critical to the ultimate fate of iron oxides and oxyhydroxides that occur as metastable phases on the Martian surface [112]. Although ferrihydrite was not one of the phases observed to form in our study, the timescales of its phase transformation provides insight into hematite formation via diagenesis on Mars. Ferrihydrite transforms in aqueous media to hematite, on timescales of years to decades, at temperatures as low as 4 • C, with the rate being primarily dependent on pH, temperature, and the presence of solutes or other minerals in the solution [113][114][115][116][117][118][119]. The extent of any diagenetic transformation is primarily a function of time and temperature [120], and mineral transformation processes studied at high temperature in the laboratory might occur during diagenesis at lower temperature after increased reaction times, unless other pathways take over. Goethite decomposes directly to hematite, through thermal or hydrothermal dehydroxylation processes [24] and by dehydration [33]. Goethite can also transform to hematite at~25 • C, during diagenesis in geological conditions, in the presence of liquid water [26]. Goethite is the most common Fe(III) bearing mineral that formed in our study ( Figure 2) and likely produces hematite, given the right conditions or time. Based on morphological [34] and spectral studies [27], goethite formed in an aqueous environment was considered to be the most probable precursor mineral for crystalline hematite spherules at Meridiani Planum [25,27,33]. Lepidocrocite, albeit a goethite polymorph, does not form hematite directly but involves maghemite (γ-Fe 2 O 3 ) as an intermediate. In our work, lepidocrocite formed in chloride-rich ( Figure 2) and perchlorate-rich solutions [49].

Implications for Gale Crater, Mars
Chlorate is likely present in Gale Crater, as suggested by in-situ SAM instrument results [51,53,54,129] and recent terrestrial laboratory experiments [55]. Chlorate is now shown to form all key iron oxides found in Gale Crater via Fe(II) oxidation (Figure 2 and [49]), except hematite, and was recently proposed to be a likely Fe(II) oxidant [130]. However, hematite is expected as a diagenetic product of many of these phases (Section 4.3). The results from the CheMin drill samples in Gale crater on Mars show coexistence of magnetite and akaganeite in the Yellowknife Bay Formation [John Klein (JK), and Cumberland (CB) drill samples] [131], and magnetite with jarosite in the Murray Formation [Confidence Hills (CH), Mojave2 (MJ), Telegraph Peak (TP), and Stoer (ST) drill samples] [6,7]. Based on the results of our studies, none of the experiments generated coexisting akaganeite and magnetite, or jarosite and magnetite. It is therefore likely that magnetite was precipitated from a fluid that was different in composition from those precipitating jarosite and akaganeite ( Figure 3). Therefore, the occurrence of hematite as well as magnetite with acidic phases indicates that multiple diagenetic fluids of different compositions have likely altered both Yellowknife Bay and Murray Formation sediments in the past. Alternatively, evaporation processes might concentrate and acidify fluids, and the same near-neutral and dilute fluid that forms magnetite might subsequently evolve into the akaganeite or jarosite stability fields, provided that enough chlorate is present to continue oxidizing residual Fe(II) and that the resulting fluids do not destroy the initial magnetite precipitates.

Conclusions
The experimental efficiency of Fe(II) oxidation by chlorate was investigated in Mars-relevant fluids. Chlorate can exhibit its maximum theoretical oxidation capacity in chloride-and sulfate-fluid environments from circumneutral to acidic pH conditions, and closely follows our predictive kinetic model [49]. The mineral products are controlled by the fluid type, acidity, and the Fe(II) oxidation rate. The oxidation rate was determined by the concentration of Fe(II), oxidant (chlorate), and the Notably, all iron oxides found in the Gale Crater can be produced by Fe(II) oxidation by chlorate or through subsequent diagenesis. Despite the low concentration on the surface of Mars (<1 wt.%), the high oxidizing capacity of chlorate and its possible regeneration by the putative oxychlorine cycle makes chlorate an important contributor of iron oxide, on both past and present Mars.
Supplementary Materials: The following are available online at http://www.mdpi.com/2075-163X/10/9/729/s1. Table S1: Mineral proportions and coherent scattering domain sizes determined from Rietveld refinements, Figure S1: [Fe(II)] and pH vs. time for control set-ups of kinetic experiments, Figure S2: Model simulations of chlorate vs. time for kinetic experiments, Figure S3: Rietveld refinement plots for minerals precipitated in Mg-chloride chlorate-equivalent experiments, Figure S4: Rietveld refinement plots for minerals precipitated in Mg-chloride chlorate-deficient experiments, Figure S5: Rietveld refinement plots for minerals precipitated in Mg-sulfate chlorate-equivalent experiments, Figure S6: Rietveld refinement plots for minerals precipitated in Mg-sulfate chlorate-deficient experiments, Figure S7: Model simulations to compare Fe(II) oxidation rate of mineral precipitation studies in relation to the minerals precipitated. The data associated with the manuscript is available at doi:10.6084/m9.figshare.12672308.