Sulfate can obscure spectral evidence of carbonate: MicrOmega observations with implications for Mars

Introduction: The Fe-carbonate siderite has been detected in Gale crater by the CheMin instrument onboard the Curiosity rover at concentrations up to ~10.5 wt% [1,2] (Fig. 1, red stars). However, it remains undetected in orbital spectral data spanning 0.4–2.5 μm, even with recent advancements in CRISM data processing [3,4]. This discrepancy raises the possibility that spectral signatures of siderite in Mt. Sharp may be masked by other minerals present in the same strata. Notably, Mg-sulfates, occurring in amorphous form as well as kieserite and starkeyite, are abundant in these layers [1–3,5] (Fig. 1, shaded pink) and exhibit strong absorptions in the ~1.9–2.4 μm range, potentially interfering with the siderite diagnostic features near 2.3 and 2.5 μm.

      Methods: We prepared five physical mixtures of Mg-sulfate and Fe-carbonate (siderite) powders in varying proportions to investigate whether Mg-sulfate can mask the orbital detection of siderite.

      MicrOmega is a hyperspectral microscope operating in the infrared range (0.99–3.65 µm), offering a 5 × 5 mm² field of view and a spatial resolution of ~22 µm per pixel [6]. It is housed within the Planetary Terrestrial Analogue Library (PTAL) chamber, a large glove box continuously flushed with pure N₂. The chamber includes a movable sample platform that enables precise positioning, as well as temperature control to manipulate the hydration state of hydrous minerals. This configuration supports repeated imaging to capture both spatial heterogeneity and temporal changes during sample dehydration. For each sample, we acquired MicrOmega hyperspectral image cubes at three timepoints: at initial chamber exposure (t₀ = 0 hours), after 3 hours (t₁), and after 1 month (t₂). We calculated key spectral parameters to aid in grain identification and examined individual spectra to identify features at both the grain and full-image scales.

Results: In line with previous studies, the Mg-sulfate in our samples progressively dehydrates when exposed to the dry MicrOmega chamber environment [7,8]. MicrOmega spectral data indicate that the Mg-sulfate is initially present as epsomite (7 H₂O per formula unit) at t₀, transitions to starkeyite (4 H₂O) after 3 hours (t₁), and becomes largely amorphous (~2 H₂O) after 1 month (t₂).

Among the diagnostic siderite features, the 2.5 μm absorption is more readily detectable than the 2.3 μm band in our mixtures. While both features are evident in spectra of individual siderite grains, only the 2.5 μm band appears in the average spectra across the full MicrOmega field of view. This suggests that in a CRISM pixel (~18 × 18 m) where Mg-sulfate is present, the 2.3 μm siderite absorption may be obscured.

Our results also show that when Mg-sulfate is in its most hydrated form—crystalline epsomite—spectral features of siderite can be almost entirely masked, even in grain-specific spectra. In contrast, less hydrated forms such as starkeyite and amorphous Mg-sulfate are less effective at obscuring siderite’s spectral signatures.

Discussion & conclusions: These findings indicate that the presence of Mg-sulfate can obscure the key 2.3 and 2.5 μm carbonate absorptions commonly used in CRISM data to identify carbonates from orbit. In Gale crater, the presence of crystalline Mg-sulfate may be a contributing factor in masking the spectral signatures of siderite. Notably, all drill sites where siderite has been identified correspond to regions mapped as polyhydrated Mg-sulfate in CRISM data (Fig. 1).

While other factors, such as the low volumetric abundance of carbonates and grain-scale textural effects, may also contribute to the absence of carbonate absorptions in orbital observations, our results suggest that hydrated Mg-sulfate in upper Mt. Sharp plays a significant role. Elsewhere on Mars, similar sulfate-rich deposits may also obscure carbonate absorptions, particularly the 2.3 μm band and, to a lesser extent, the 2.5 μm feature, when carbonates are present at ~<25 wt%. This is relevant to other works that propose that major layered sulfate units could contain carbonates at levels comparable to those in Gale crater [2].

In sulfate-bearing terrains, particular attention should be given to the 2.5 μm band, as it may offer the only visible indication of carbonates. This is especially relevant for sites like Oxia Planum, where carbonate detections rely solely on a 2.5 μm feature without a corresponding 2.3 μm band [13]. Revisiting surface spectra to identify isolated 2.5 μm absorptions may provide new insights, especially considering the widespread nature of secondary sulfates on Mars and their impact on orbital carbonate detection.

 

References: [1] Thorpe et al., 2022, JGR Planets, 127, e2021JE007099. [2] Tutolo et al., 2025, Science, 388, 6744. [3] Sheppard et al., 2020, JGR Planets, 126. [4] Dhoundiyal et al., 2023, Icarus, 115504. [5] Fraeman et al., 2016, JGR Planets, 121. [6] Bibring et al., 2017, Astrobiology, 17, 621-626. [7] Sheppard et al., 2022, Icarus, 115083. [8] Chou et al., 2013, Journal of Asian Earth Sciences, 62. [9] Viviano-Beck et al., 2014, JGR Planets, 119. [10] Chipera et al., 2023, JGR Planets, 128. [11] Loizeau et al., 2020, Planetary and Space Science. [12] Farrand et al., 2024, AGU abstract.   [13] Mandon, L., Parkes Bowen, A., Quantin-Nataf, C., Bridges, J. C., Carter, J., Pan, L., et al. 2021. Astrobiology, 21, 464–480.

 

Fig. 1: Orbital map of Curiosity drill holes. Samples where Curiosity detected Fe(II)-carbonate with CheMin [1] are marked with red stars and their abbreviated drill names: Glen Etive (GE/GE2), Mary Anning (MA/MA3), Groken (GR), Nontron (NT), Bardou (BD), Tapo Caparo (TC), Ubajara (UB), Sequoia (SQ). Samples where no carbonate was detected are marked with white circles. The drill holes are superimposed over HiRISE imagery and show in pink where orbital CRISM spectra show the presence of polyhydrated Mg-sulfate [3].