Siderite and ferric oxyhydroxides imply interlinked carbon, iron, and halogen cycles on Mars

Significance

The overall rarity of carbonate deposits on Mars have been interpreted to be the result of acid dissolution. We show that siderite is stable in Mars-relevant fluids from pH 7 to 2 and its alteration to ferric minerals—similar to those recently discovered at Gale crater—is not possible by acidic diagenesis alone. Our study reveals that chlorate and bromate, the dominant form of chlorine and bromine on Mars, can oxidatively weather siderite to form ferric hydroxide minerals observed on Mars. Siderite weathering by oxyhalogen brines not only operates in acidic solutions but is also effective at near-neutral conditions.

Abstract

Pure siderite [Fe^IICO₃] was recently discovered in abundant quantities (4.8 to 10.5 wt.%) by the Curiosity rover at Gale crater, Mars. Diagenetic alteration of siderite likely caused the carbonate-sequestered CO₂ to be released back into the atmosphere and consequently produced ferric [Fe(III)] oxyhydr(oxide) minerals. Here, using laboratory experimentation, we demonstrate that while closed system acid diagenesis—as proposed for Gale crater—is incapable of effective siderite alteration in Mars-relevant fluids, oxyhalogen compounds (chlorate and bromate) can weather siderite not only at acidic pH but also in near-neutral Mars-relevant solutions. The ferric oxyhydroxide minerals produced as a consequence are controlled by the diagenetic fluid composition. While photooxidation is possible, the mutually exclusive products of alteration—magnetite (Fe₃O₄) during ultraviolet irradiation and ferric oxyhydroxide (FeOOH) by oxyhalogens—demonstrate that siderite at Gale crater underwent chemical weathering by chlorate and bromate brines owing to the complete absence of magnetite in drill samples containing siderite. We propose a top–down oxyhalogen brine percolation model to explain the iron mineralogy of the sulfate-rich unit at Gale crater. We conclude that siderite alteration by acidic fluids alone cannot explain the redox disequilibrium witnessed in Gale crater sediments as promulgated before and siderite weathering by oxyhalogen brines is the most likely explanation. It is highly likely that the halogen cycle on Mars is interlinked to the iron and the carbon cycle on early and current Mars.

The current Martian atmosphere is carbon dioxide (CO₂)-rich but thin (P_atm ~ 6 mbar) (1). The early Martian atmosphere likely had a thicker CO₂-rich atmosphere (pCO₂ ~ 1.5 bar) (2). Geochemical processes (e.g., silicate-carbonate cycle, geological carbon cycle) can shift the CO₂ inventory from the atmosphere to the crust by sequestering it as carbonate (CO₃^2-) minerals (3–6). But large-scale carbonate deposits remained obscure on Mars and extensive acidic dissolution was proposed to reconcile with these “missing” carbonates (7, 8).

Recently, crystalline ferrous carbonate mineral siderite [Fe^IICO₃] was discovered at Gale crater (9). Abundant (as high as 10 wt.%) highly pure siderite—as opposed to other detections with metal substitutions (10, 11)—was detected over 89 m of the Mg-sulfate-rich Chenapau member of the Mirador formation, sequestering about 2.6 to 5.7 mbar of CO₂. Siderite formation in this sulfate-rich deposit, however, is not representative or directly caused by an elevated atmospheric CO₂ (9). Rather, this pure siderite-rich unit represents a major transition event in the sedimentary cycle on Mars governed primarily by surface geochemistry. Siderite saturated as Fe²⁺ and (bi)carbonate-rich anoxic fluids underwent evaporative concentration (or freezing) in the subsurface pore spaces that are cut off from the environment (6, 9, 12). Upon reaching their respective saturation points, siderite and even soluble Mg-sulfate salts, precipitated in this closed system environment that is representative of the eventual “drying” of the planet.

While other iron-bearing minerals—hematite [Fe^III₂O₃], goethite [α-Fe^IIIOOH], and akaganeite [β-Fe^IIIO(OH, Cl)]—were detected along with siderite [Fe^IICO₃], they did not form by evaporative concentration of anoxic Fe(II)-rich solutions. Rather, postdepositional diagenetic alteration of ferrous carbonate to ferric oxyhydr(oxide) minerals is responsible for the juxtaposition of both Fe(II)- and Fe(III)-bearing minerals. An acid alteration hypothesis has been invoked to explain this observation in Gale sediments (9, 13) but is primarily based on terrestrial [e.g., (14, 15)] and meteoritic (16) studies. Extensive literature on siderite dissolution under various aqueous and atmospheric conditions exist but fail to accurately represent the geochemical condition of the recent siderite discovery at Gale crater. While most studies on siderite dissolution focus on its behavior in pure water or dilute solutions exposed to an O₂-rich atmosphere, conditions relevant primarily to Earth (supplementary section), we investigate siderite dissolution under conditions tailored to Mars. By doing so, our work explores how siderite alters in a setting far removed from Earth’s oxygen-rich, water-saturated norm, offering insights into iron carbonate stability and its role in Mars’ geological history.

Limited Siderite Alteration in Oxidant-Free Martian Fluids

Siderite alteration at Gale crater likely occurred in concentrated Mg-sulfate- or Mg-chloride-solutions in closed systems that were decoupled from the surface, thereby rendering the effect of atmospheric composition inconsequential (9). Batch reactors simulating siderite weathering in Gale crater subsurface by Mars-relevant fluids (MgCl₂- and MgSO₄-rich solutions between pH 7 and 2) demonstrated minimal dissolution and chemical alteration after ~100 days of laboratory experimentation (Fig. 1 and Table 1) (see Materials and Methods for details). Siderite did not experience extensive acid diagenesis—contradicting previous hypothesis (8)—but instead remained almost unaltered and settled at the bottom of the reactors with clear solutions. All solutions experienced an increase in pH from ~2.5, 4, and 7 to about 5, 7, and 7.5, respectively. X-ray diffraction (XRD) patterns of siderite in control experiments did not produce any new peaks but experienced minor decrease in siderite peak intensities and increase in background intensity in both acidic and near-neutral control solutions as compared to untreated siderite (Fig. 1).

Fig. 1.

The characterization of siderite before and after being subjected to alteration in the control experiments without oxyhalogen salts. (A) The XRD scan of the untreated siderite and Rietveld refinement of the corresponding XRD pattern. (The calculated parameters from Rietveld refinement on XRD scans presented in the first row of SI Appendix, Table S5) (B) Scanning electron microscopy (SEM) images of the untreated synthetic siderite. (C–F) The XRD patterns of altered siderite (FeCO₃) in control experiments without the presence of oxidants in background fluid containing MgCl₂ (C and E) and MgSO₄ (D and F) after 35 d (C and D) and 100 d (E and F). The calculated Rietveld refined XRD scans (in red) are overlain on the observed scans (in blue). The sample labels alongside each scan depict the fluid type (Cl = Mg-chloride fluid, S = Mg-sulfate fluid), Ctrl = control reactors, and the initial pH (further sample details in Table 1). The peak labels: Sd = siderite, Sd* = minor impurities in the powdered siderite sample; Star marks = remnant salt in alteration assemblage. (G) The photographic image of the control reactors at the end of the 100 d experiments in 100 mmol L⁻¹ MgCl₂ background fluid (H) and 100 mmol L⁻¹ MgSO₄ background fluid.

Table 1.

Fluid composition of long-term (100 d) siderite alteration experiments containing ~100 mmol L⁻¹ oxidant (ClO₃⁻ or BrO₃⁻) and control experiments

Sample*	Initial pH	Final pH†	Initial [ClO3−] or [BrO3−]	Final [ClO3−] or [BrO3−]‡	Total Dissolved Fe (mmol L−1)c
Control Experiments without chlorate or bromate; Background fluid containing ~100 mmol L−1 MgSO4
SidLong9 [Ctrl-S-pH3]	2.63	4.99	N/A	N/A	1.49
SidLong10 [Ctrl-S-pH5]	4.40	6.67	N/A	N/A	0.11
SidLong11 [Ctrl-S-pH7]	6.47	7.34	N/A	N/A	0.25
Control Experiments without chlorate or bromate; Background fluid containing ~100 mmol L−1 MgCl2
SidLong21 [Ctrl-Cl-pH3]	2.20	4.87	N/A	N/A	1.86
SidLong22 [Ctrl-Cl-pH5]	3.47	6.84	N/A	N/A	0.24
SidLong23 [Ctrl-Cl-pH7]	7.01	7.48	N/A	N/A	0.01
Sample *	Initial pH	Final pH †	Initial [ClO 3 − ] (mmol L −1 )	Final [ClO 3 − ] (mmol L −1 ) ]‡	Total Dissolved Fe (mmol L −1 ) c
Background fluid containing ∼100 mmol L−1 MgSO4
SidLong5: S-pH3-ClO3	2.96	4.85	100	91	0.19
SidLong6: S-pH5- ClO3	4.35	6.15	92	94	0.15
SidLong7: S-pH7-ClO3	6.39	7.45	101	108	0.04
Background fluid containing ~100 mmol L−1 MgCl2
SidLong17: Cl-pH3-ClO3	2.36	4.67	145	X§	0.14
SidLong18: Cl-pH5- ClO3	3.33	6.11	90	X§	0.01
SidLong19: Cl-pH7-ClO3	6.59	7.21	109	X§	0.12
Sample *	Initial pH	Final pH †	Initial [BrO 3 − ] (mmol L −1 )	Final [BrO3−] (mmol L−1)‡	Total Dissolved Fe (mmol L −1 ) c
Background fluid containing ~100 mmol L−1 MgSO4
SidLong1: S-pH3-BrO3	2.57	4.15	99	67	0.01
SidLong2: S-pH5-BrO3	4.79	6.20	96	106	0.05
SidLong3: S-pH7-BrO3	6.84	6.05	100	107	0.04
Background fluid containing ~100 mmol L−1 MgCl2
SidLong13: Cl-pH3-BrO3	2.40	4.59	100	67	0.03
SidLong14: Cl-pH5-BrO3	3.77	5.64	100	63	0.04
SidLong15: Cl-pH7-BrO3	7.03	6.04	119	64	0.11

^*Sample name code.

^†pH and oxidant after ~100 d of start of the experiments.

^‡Concentration after ~100 d of start of the experiments.

^§Error while gathering data.

The results from batch reactor control experimentation demonstrate the resistance of siderite to undergo dissolution in batch reactors in Mars-relevant fluids from pH ~7 to 2. The control solutions quickly saturate with respect to siderite and reach near-equilibrium condition that disallows any further dissolution in Mars-relevant fluids in a closed system (SI Appendix). While one might argue that the initial H⁺ ions may be enough for complete siderite dissolution, it is unlikely for subsurface pore spaces on “drier Mars” to have acidic (pH < 2) groundwater in abundance greater than our experiments (water to rock ratios~250). Therefore, substantial siderite alteration by acidic fluids in closed, water-limited Martian subsurface condition alone is highly unlikely. Siderite is much more stable in Mars-relevant fluids and current calculations on siderite dissolution and alteration by acidic fluids on Mars are likely overestimates (17). Another possible pathway for siderite alteration involves exposure to ultraviolet (UV) radiation, triggering photochemical reactions that oxidize Fe(II) in siderite to Fe(III) and consequently produce magnetite (Fe₃O₄) in anoxic aqueous solutions (18). However, siderite discovered at Gale crater does not appear to have undergone significant UV-induced alteration, likely due to the limited penetration of UV irradiation and the complete absence of magnetite in the drill samples containing siderite (see SI Appendix). For siderite to undergo alteration and produce ferric minerals, stronger oxidative agents capable of accessing siderite filled pore spaces at Gale crater are required. Can “oxic” brines—as earlier proposed to be an important oxidant (19) on Mars—cause siderite diagenesis at Gale crater?

The most common oxic brines on Mars are composed of oxychlorine salts (20–22). Oxychlorine salts, like chlorate (ClO₃^-), are effective iron oxidants under Mars-relevant conditions (23–29), comprise a substantial proportion of total surficial chlorine (Cl), and are globally distributed on the planet (30–32). Multiple orbiter and rover instruments have identified (per)chlorate salts in Martian soil, sediment, rock, and meteorite samples (22, 31, 33–35). Altogether, an active oxychlorine cycle in the past (27, 36) and a greater production in the present thin atmosphere on Mars (20, 21, 37, 38) have been implied. Owing to their extremely low eutectic temperatures (34), oxychlorine brines have been proposed to stabilize on the current Martian surface on a regular diurnal cycle (39). Oxychlorine salts thus form excellent oxidizing brines. In addition to Cl, Mars is also enriched in bromine (Br), about four times as compared to Earth (40, 41). All samples analyzed by Spirit, Opportunity, and Curiosity rovers (41–44) have detected Br and laboratory studies have found bromate [BrO₃^-] to be the most likely Br phase on the surface of Mars (45). While the effect of chlorate and bromate on Fe(II) bearing minerals like iron sulfides (24) and magnetite (29) have been studied, siderite alteration by chlorate and bromate in Mars-relevant fluids under closed system conditions remain uninvestigated. Here, using batch experiments in the laboratory, we recreate Gale pore space conditions to investigate the extent of siderite weathering in Mars-relevant oxyhalogen brines and analyze the resultant alteration mineral assemblage.

Results

Siderite Alteration by Chlorate.

Laboratory experiments were conducted to investigate the weathering of siderite by oxyhalogen brines in different Mars-relevant fluids at ambient conditions (24 °C, 1 atm [97% N₂ and 3% H₂]). The experiments were designed to evaluate the oxidative weathering products of siderite as a function of oxyhalogen type [sodium chlorate (NaClO₃) and sodium bromate (NaBrO₃)], background fluid type [magnesium chloride (MgCl₂) and magnesium sulfate (MgSO₄)], initial solution pH [~2.5, 4.5, and 7] and time [35 and 100 d] (see Materials and Methods for details).

Within 35 d of experimentation, chlorate altered siderite to produce goethite (SI Appendix, Figs. S2 and S3). While XRD peak intensities of siderite decreased in all reactors, the greatest reduction was seen in acidic solutions. This implies that siderite alteration intensity increased with acidity. In acidic Mg-chloride and Mg-sulfate solutions, broad but prominent goethite peaks appeared amounting to about 15 ± 6 and 20 ± 3 wt.% of the product, respectively, as determined by Rietveld refinement of XRD data. Similarly, in mildly acidic Mg-chloride and Mg-sulfate solutions, 7 ± 5 and 9 ± 5 wt.% goethite was produced. Signs of siderite grain alteration were evident in SEM images. While typical acicular morphology of goethite was not detected, the otherwise smooth surface of siderite is pockmarked and roughened due to the dissolution, alteration, and the growth of nanoparticulate goethite.

After 100 d of experimentation, siderite in all chlorate-containing reactors underwent oxidative weathering and produced typical orange-yellow hue that is characteristic of ferric minerals (SI Appendix, Fig. S1). The extent of siderite alteration and the resulting mineral alteration assemblage is a strong function of solution pH and background fluid. Siderite underwent greater dissolution in acidic reactors, showed an increase in solution pH from ~2.5 to ~5, and generated higher concentrations of dissolved iron than experiments at higher initial pH (Table 1 and Fig. 2 A and B). Siderite alteration in acidic Mg-chloride fluids produced a combination of goethite (41 ± 6 wt.%) and akageneite [β-FeO(OH,Cl)] (20 ± 6 wt.%); the Mg-sulfate analog yielded only goethite (50 ± 6 wt%). Lepidocrocite [γ-FeOOH] (11± 3 wt.%) was produced in mildly acidic Mg-sulfate solutions. What is critical to note here is that not only did chlorate promote siderite alteration in acidic fluids, but also produced ~12 wt.% goethite in near-neutral solutions (~pH 6.5) in both Mg-chloride and Mg-sulfate fluids.

Fig. 2.

Characterization of siderite altered by sodium chlorate (A–F) and sodium bromate (G–K) in Mars-relevant fluids. XRD patterns (A and G), and the mineral assemblage obtained after conducting Rietveld refinement of XRD patterns (B and H), SEM images (C–F and I), and (J and K) transmission electron spectroscopy (TEM) images of siderite (FeCO₃) alteration experiments at the end of the 100 d experiments in presence of 100 mmol L⁻¹ sodium chlorate (NaClO₃) (A–F) and 100 mmol L⁻¹ sodium bromate (NaBrO₃) (G–K) in background fluid containing MgCl₂ and MgSO₄ after 100 d. The sample labels alongside each scan depict the fluid type (Cl = Mg-chloride fluid, S = Mg-sulfate fluid), the initial pH, and the oxidant (ClO₃ = chlorate and BrO₃ = bromate) (further sample details in Table 1). The peak labels: S = siderite, S* = minor impurities in siderite sample, A = akaganeite, G = goethite, L = Lepidocrocite.

Goethite needles are seen growing on siderite grains and spherical clusters of acicular goethite are strewn around in the field of view of typical SEM images (Fig. 2 C–F). The surface of siderite grains is more altered, and goethite is seen pervasively on the outer surface of siderite grains while maintaining the morphology of the original siderite grain. Siderite altered by chlorate at pH <~4.5 show similar evidence of acicular goethite particles growing directly on carbonate grains. However, the siderite sample at near-neutral pH did not show any evidence of acicular goethite particles growth on siderite grains but showed surface alteration features similar to those observed at acidic solutions in short-term experiments (SI Appendix, Fig. S3).

Siderite Alteration by Bromate.

Within 35 days, bromate altered siderite and produced goethite in all experimental solutions irrespective of background fluid composition or pH (SI Appendix, Fig. S4). The extent of siderite alteration and the resulting formation of goethite increased with increasing fluid acidity. About 15 ± 6, 23 ± 8, and 29 ± 5 wt.% goethite formed by siderite alteration in MgSO₄ fluid at initial ~pH 6.7, 4.1, 2.4, respectively. Analogous reactors in MgCl₂ fluids produced 15 ± 4, 17 ± 5, and 13 ± 4 wt.% goethite at ~pH 7.3, 4.3, 3.3, respectively. SEM images revealed striking contrast between the grain size and crystal morphology between siderite and the resultant alteration products (SI Appendix, Fig. S5). The alteration products are seen growing on the smooth surface of siderite leading to the generation of an overall weathered texture on siderite grains in all experimental samples.

After 100 d, siderite underwent substantial oxidative alteration and produced different ferric oxyhydroxide minerals in both Mg-chloride and Mg-sulfate fluids across acidic to near-neutral fluids (Fig. 2 G–K). The final pH of reactors at initial acidic and mildly acidic solutions increased by ~2 pH units while near-neutral solutions experienced a drop in pH by ~1 pH unit in both Mg-chloride and Mg-sulfate fluids (Table 1). Most importantly, siderite experienced substantial oxidative alteration in near-neutral fluids (34 to 50%) that increased (~65%) in acidic fluids and produced goethite in both MgCl₂ and MgSO₄ fluids across ~pH 2.5 to 7. About 23 ± 14 wt.% akageneite along with 33 ± 13 wt.% nanoparticulate goethite was produced by siderite alteration in acidic Mg-chloride fluids. The Mg-chloride fluids at ~pH 4 and ~pH 7 produced 44 ± 7 and 48 ± 5 wt.% goethite along with 1 ± 0.8 and 3 ± 2 wt.% akageneite, respectively. In contrast, MgSO₄ fluids mainly produced 42 ± 8, 25 ± 5, and 34 ± 8 wt.% goethite at initial pH 2.6, 4.8, and 6.8, respectively. Minor amount of lepidocrocite (1 ± 0.5 wt.%) formed in near-neutral MgSO₄ fluids.

Discussion

The Extent of Alteration Is Determined by Reaction Timescales and Diagenetic Fluid Type.

Greater amounts of siderite underwent alteration in the 100 d long-term experiments than those within 35 d thereby producing a greater amount of alteration mineral (SI Appendix, Fig. S6). Additionally, a more diversified alteration mineralogy was observed at lower pH in the long-term experiments. The relative increase in siderite alteration with the passage of time can be captured by SEM images. While small sand-like particles of goethite are found only partially coating siderite grains in short-term experiments (e.g., SI Appendix, Fig. S5F), they completely cover siderite grains in the long-term (Fig. 2). Longer duration experiments are likely to increase siderite alteration as oxyhalogens remain available in experimental solutions (Table 1), demonstrating the weathering effectiveness of oxyhalogen compounds.

Siderite underwent progressively higher amounts of oxidative weathering by oxyhalogens with a decrease in pH from near-neutral to acidic conditions. While chlorate at near-neutral pH resulted in ~10% siderite alteration, acidic solutions caused as much as 60% alteration in acidic conditions. Similarly, siderite alteration by bromate increased from 34 to 66% when initial pH decreased from ~7 to 2.5 (SI Appendix, Fig. S6). Also, bromate is more (2 to 3×) effective than chlorate in oxidizing siderite at near-neutral and mildly acidic fluids as expected from its greater standard redox potential (SI Appendix, Table S6). In acidic fluids, both chlorate and bromate are almost equally efficient in weathering siderite. The amount of siderite alteration is dependent more on solution pH and oxidant while less on background fluid type. The influence of background fluid on the amount of siderite alteration is complicated but its effect on the type of mineral product is more straightforward.

Siderite Alteration by Oxyhalogen Brines Produces Ferric Oxyhydroxide Minerals.

While dissolved Fe(II) (23, 25, 26) and ferrous sulfide minerals (24) oxidation have previously been shown to produce a diversified alteration assemblage, siderite alteration by oxyhalogens primarily formed goethite. All short-term experiments—irrespective of their pH, oxidant, or background fluid—exclusively produced goethite (SI Appendix, Figs. S2 and S4). Morphological changes were observed, with chlorate being the most effective alteration agent, producing goethite, identifiable by its needle-like structure in SEM images. In contrast, bromate generated ~1 µm-sized particles covering the outer surface of siderite, which displayed distinct shapes different from the needle-like goethite; TEM images showed growth of goethite needles on siderite by bromate alteration (Fig. 2J). The amount of goethite produced is directly proportional to the extent of siderite alteration, and therefore strongly influenced by solution pH and oxidant type (see section above). In long-term experiments, goethite remained as the primary, and in most cases the only, alteration product (SI Appendix, Fig. S7). But other ferric oxyhydroxide minerals, akaganeite and lepidocrocite, co-occur with goethite in specific experimental conditions governed by solution chemistry. To reiterate, siderite alteration by oxyhalogen brines exclusively produced ferric oxyhydroxides and not magnetite that were produced during siderite photooxidation (18). Hence, the ferric minerals at Gale crater are likely an outcome of siderite alteration by chlorate and bromate brines.

The effect of background fluid on alteration product is complicated. While the formation of goethite [α-FeOOH] does not show any preference to background fluid, lepidocrocite [γ-FeOOH] and akaganeite [β-FeO(OH,Cl)] precipitate exclusively in sulfate- and chloride-rich solutions, respectively. Siderite initially altered to form goethite and then likely transformed to its polymorphs depending upon the diagenetic fluid composition. Contrary to previous observation (23, 46), lepidocrocite precipitated in sulfate- rather than chloride-rich fluids during siderite alteration. Akaganeite (~20 wt.%) formed exclusively in acidic chloride-rich fluids by both chlorate and bromate at initial ~pH 2.40. Interestingly, minor amounts of akaganeite (1 to 3 wt.%) also formed in mildly acidic to near-neutral chloride-rich solutions containing bromate. To our knowledge, no other process has shown akaganeite formation at near-neutral (~pH 7.03) conditions during siderite alteration. Our results, therefore, demonstrate an important formation pathway for akageneite that is not restricted to acidic pH.

The Carbon–Iron–Halogen Cycle on Mars Is Interlinked.

The Martian carbon cycle is intricately connected to the sedimentary iron cycle, as ferrous carbonate can act as a sink as well as a source of CO₂. Siderite destruction released CO₂ that was once trapped in the carbonate mineral, back to the atmosphere and thereby “closing the loop” of the carbon cycle on Mars (9). Minerals related to the stratigraphic sections containing abundant siderite mark a substantial and abrupt shift in geochemical processes occurring at Gale crater. The Mg-sulfate-bearing unit above the Canaima (CA) drill site (−3,879 m elevation) signifies the beginning of a drier climate on Mars (47). The CA drill site, just ~26 m below the most siderite-rich unit, does not show any sign of carbonate mineral but contains ferric oxyhydr(oxide) minerals. The overlying Amapari Marker Band (AMB) (−3,853 ± 1 m) is a visually and geochemically distinct, erosion-resistant layer that is substantially metal [iron, manganese, chromium, zinc] rich and presents the highest concentrations of iron (Fe) (47 wt.%) ever recorded by the APXS instrument (48). Interestingly, the same metal-rich AMB records the highest concentrations of chlorine (3 wt.%) and bromine (0.44 wt.%) during the “northern crossing” investigation of the Curiosity rover.

The Tapo Caparo (TC) drill in the overlying thick-laminated strata (−3,853 m) revealed the highest ever recorded abundance of crystalline pure siderite (10.5 ± 0.5 wt.%) but no evidence of Fe(III) minerals (9). The Ubajara (UB) drill sample (−3,826 m) up-section contained 4.8 ± 0.3 wt.% siderite and 3.5 ± 0.7 wt.% ferric (oxyhydr)oxide; The Sequoia (SQ) drill sample (−3,764 m) further up contained 7.6 ± 0.4 wt.% siderite and 5.1 ± 0.7 wt.% ferric (oxyhydr)oxide. Only the topmost SQ drill sample contains akaganeite (2.7 ± 0.6 wt.%) that requires a Cl-rich acidic fluid for its formation (46) (Fig. 3). The geochemical processes that eventually lead to the unique mineralogy of the region holds the key to the carbon and iron cycles on Mars.

Fig. 3.

A schematic of the “top–down oxyhalogen brine percolation model” representing the Gale crater stratigraphy and the interconnected iron, halogen, and carbon cycle proposed on Mars. Oxyhalogen salts form and accumulate on the surface of Mars that form dense oxic brines that percolate downward due to the gravity. Near the top of the stratigraphy (close to drill sample SQ) the oxyhalogen brines are likely acidic and Mg-chloride rich. They react with sedimentary siderite and likely produce an alteration mixture assemblage of akaganeite and goethite. The oxyhalogen brine undergoes neutralization and oxidatively weather siderite to only goethite (close to drill sample UB). Both goethite and akaganeite can undergo diagenetic alteration to hematite observed at Gale crater. By the time the brine reaches the bottom of the stratigraphy close to drill sample TC, all chlorate and bromate has been consumed and siderite does not undergo any alteration. The resultant chloride and bromide salts accumulate and get collected at the putatively impermeable Amapari Marker Band. The oxyhalogen brines therefore alter siderite to produce the iron mineralogy depicted by the pie chart and release the carbonate sequestered CO₂ back to the atmosphere. The estimates from mass balance calculations represented inside dashed boxes.

The simultaneous discovery of sediments containing the highest ever recorded siderite, chlorine, and bromine, along with a downward decreasing ferric mineral (akaganeite + goethite + hematite) trend is most likely related in the stratigraphic section in and above impenetrable and likely impermeable AMB. We propose that a downward percolating oxyhalogen-rich brine in the Mg-sulfate-rich unit at Gale crater that prompted a top–down oxidative alteration of siderite, formed ferric oxyhydroxide minerals as a consequence, and released carbonate-trapped CO₂ back to the atmosphere (Fig. 3).

An acidic Mg-chloride brine containing chlorate or bromate percolated into the siderite-rich strata from the top, oxidatively weathered it, and produced goethite and akaganeite (SQ drill site). We note that even a near-neutral bromate brine may be responsible for siderite alteration to akaganeite. The brine migrated further deeper down-section and weathered siderite to form only goethite as brine acidity and oxidant concentration decrease (UB drill site). By the time the brine reaches the lower levels, all chlorate and bromate has been consumed and siderite remains unaltered (TC drill site). At this elevation, chlorate and/or bromate is fully reduced to chloride (Cl⁻) and bromide (Br⁻) that is trapped by the impermeable AMB. About 7 × 10¹⁵ kg sodium chlorate is required to alter siderite in the upper ~62 m of the siderite bearing section between drill holes SQ and UB (see supplementary section). About 3.7 × 10¹⁵ kg sodium chlorate—half of the amount required—is likely available in the stratigraphic section assuming ~1 wt.% chlorate distribution in Gale crater sediments (22). The presence of bromate in the sediments is likely responsible for the alteration of the other half of the available siderite at Gale crater. Oxyhalogen brines could also lead to oxidative weathering of siderite elsewhere and could lead to their disappearance globally on the planet and aid to climate change by effectively releasing trapped CO₂ back to the atmosphere. Owing to their distribution, abundance, extremely low eutectic temperatures, and a conducive modern atmosphere, it is highly likely that siderite alteration by oxyhalogen brines is an ongoing process on modern Mars. A complex, active halogen cycle has recently been proposed on Early Mars (27) and is likely intricately tied to both carbon and iron cycle on modern “drier” Mars.

Materials and Methods

Experimental Design.

Laboratory experiments were conducted to investigate the weathering of siderite by oxyhalogen brines in different Mars-relevant fluids at ambient conditions (24 °C, 1 atm). The experiments were designed to evaluate the oxidative weathering products of siderite as a function of oxyhalogen type [sodium chlorate (NaClO₃) and sodium bromate (NaBrO₃)], background fluid type [magnesium chloride (MgCl₂) and magnesium sulfate (MgSO₄)], initial solution pH [~2.5, 4.5, and 7] and time [35 and 100 d]. Magnesium sulfate and chloride salts were specifically chosen owing to their prevalence on Mars (49, 50) and to match experimental conditions to past studies (17, 24, 29). The background salts serve as ionic strength buffers and provide anions with different complexation behavior toward dissolved Fe(II/III), thereby likely affecting reaction rates and the mineral products of siderite dissolution and alteration (17).

Experimental Setup.

All experiments in the study were conducted inside a Coy vinyl anaerobic chamber (N₂ = 97%, H₂ = 3%; O₂ < 1 ppmv maintained using palladium catalysts). The 50 mL glass serum bottle reactors used in the experiments were wrapped in Al foil to prevent any inadvertent influence of ultraviolet radiation on siderite dissolution and oxidation. Deionized ultrapure (18.2 MΩ cm) (DI) water for the experiments was deoxygenated outside the anaerobic chamber by bubbling ultrapure N₂ into a 1 L Erlenmeyer flask at 75 °C for three hours; the flask was quickly transferred to the anaerobic chamber, and checked for dissolved O₂ colorimetrically (CHEMetrics test kit K-7540). The DI water used in the experiments contained dissolved oxygen below the detection limit of 2.5 μg L⁻¹ (~80 nmol L⁻¹). Fisher Scientific ACS-grade reagents were dissolved in deoxygenated DI water to prepare 2 mol L⁻¹ stock solutions of sodium chlorate [NaClO₃], sodium bromate [NaBrO₃], magnesium chloride hexahydrate [MgCl₂⋅6H₂O], magnesium sulfate hexahydrate [MgSO₄⋅6H₂O] and mixed in appropriate amounts to achieve target experimental compositions.

Siderite [Fe^IICO₃] for the experiments was procured from Strem Chemicals Inc. (MA) through Fisher Scientific. The technical grade sample was delivered to the laboratory in sealed, N₂-flushed glass bottle; it was opened and contained in an anoxic environment for the entirety of the experiments. The untreated sample was characterized prior to the experiments using XRD and SEM (SI Appendix, Fig. S1).

Types of Experiments.

Three types of experiments in different Mars-relevant fluids were designed. The first was a set of control experiments without oxidants that was run for ~100 d. The second and third set were solutions containing (~100 mmol L⁻¹) chlorate- and bromate, respectively, that were run for a short-term (35 d) and a long-term (100 d). 50 mL solutions containing 0.2 g siderite (~4 g L⁻¹; W:R ~ 250) were allowed to react with ~100 mmol L⁻¹ NaClO₃ or NaBrO₃ in background salt mixtures of 100 mmol L⁻¹ Mg-chloride or Mg-sulfate at initial ~pH 2.5, 4, and 7. 12 reactors each were set for the control, chlorate-containing, and bromate-containing solutions for the long term (SI Appendix, Table S1). Another set with 12 reactors each were set up for chlorate and bromate for the short-term experiments (SI Appendix, Table S4). Our specific experimental conditions aim to mimic (sub)surface alteration conditions at Gale crater region and allow us a comparative analysis with other studies (24).

The experiments were initiated by adding DI water, oxidant (BrO₃⁻ or ClO₃⁻), and background fluid (MgSO₄ or MgCl₂) in accurate proportions in the reactors. The initial pH of the experimental solutions was set to ~7, 4, or 2.5 by adding 1 mol L⁻¹ solutions of hydrogen chloride or sodium hydroxide using a Thermo Scientific Orion Star pH A211 pH meter with Orion ROSS Ultra Triode (gel-filled, pH/ATC electrode). Siderite was weighed and added as the final component to the reactors, following which the pH was allowed to drift freely in response to putative siderite dissolution and Fe(II) oxidation. The reactors were capped with blue butyl stoppers and sealed with Al crimpers to avoid inadvertent oxidation by other oxidants (e.g., stray light, O₂) inside the anaerobic chambers. Sealed reactors were brought out of the anaerobic chamber and placed in drawers in the dark without being stirred or agitated during the entire duration of the experiments in order to replicate a stagnant surface/pore water environment, similar to an open lake or a pore-spaces in the Martian subsurface.

At the end of the experiments, all reactors were brought inside the anaerobic chamber and filtered using a 0.22 μm pore size MCE membrane. The filtrates were washed with DI water to remove excess salts from the alteration assemblage. The filtrates were dried inside the anaerobic chamber in a vacuum desiccator at 24 °C. Since bromate reduction leads to the production of toxic bromine gas (27), the reactors containing bromate at initial pH 2.5 were unsealed outside the anaerobic chamber and kept overnight in a fume hood to allow release of dissolved Br₂ gas. The final solution pH was measured at the end of each experiment. The filtrates and the supernatant fluids were analyzed by a suite of instruments at the University of Texas at San Antonio and Washington University that included XRD, SEM, TEM, inductively coupled plasma optical emission spectrometer, and ion chromatography (see SI Appendix for details).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All analytical datasets related to this research is publicly available (https://figshare.com/articles/dataset/dx_doi_org_10_6084_m9_figshare_27225450/27225450) (51). All other data are included in the manuscript and/or SI Appendix.