Towards new methods for identifying hydrated sulfates on Mars using Visible-Infrared (VISIR) spectroscopy

Hydrated sulfate minerals on Mars are of critical importance to understand the ancient climates of the planet, as they can form through multiple processes involving water-rock interaction. Particularly, they can form as evaporitic deposits ^[1], when bodies of liquid water containing dissolved salts slowly evaporate, transitioning to brine-like environments and finally leaving behind the crystallized salts they once contained. This makes them ideal to trace ancient aqueous environments. These salts have been detected in several places on Mars ^[2,3,4], but are not as widespread as initially expected. Additionally, they are rarely seen alongside the other phases they have been modelled to form alteration sequences with, namely the clay minerals dominating the Southern Noachian terrains ^[5]. This could be explained by some geological processes, but also by incomplete detections and/or observational biases.

To identify mono- and poly-hydrated sulfates (MHS and PHS respectively), one technique available both orbitally and in-situ is Visible and Near-Infrared (VISIR) reflectance spectroscopy. But even though hydrated sulfate minerals present distinctive spectral signatures, their identification is rarely straightforward. Indeed, other families of alteration minerals also detected on Mars ^[2,3,4] (such as phyllosilicates, carbonates, amorphous hydrated silica, …) bear spectral signatures in the wavelength ranges available in current Martian data (typically from 1.3 to 2.6 µm) ^[6,7] that can overlap with sulfate spectral features. This problem becomes exacerbated in the case of mineral mixtures, where spectral contributions from several phases make clear identification challenging ^[8,9]. Other studies have come up with spectral parameters capable of detecting hydrated sulfates ^[10], and while they can work well in some conditions, they can also mix them up with other mineralogical families.

Here, we develop a new suite of spectral parameters capable of confidently identifying MHS and PHS using VISIR data. The objective is to automatically detect hydrated sulfate-bearing outcrops with minimal ambiguity, even when mixed with other phases. To do so, we base our study on 2 key absorption features in hydrated sulfate spectra: (1) an important drop in reflectance near 2.4 µm (S-O overtone) ^[11], present in other hydrated minerals but usually weaker in intensity; and (2) for most sulfates, a spectral plateau lacking any absorption feature in the 2.2-2.3 µm range. Features in that range may be linked to combinations/overtones of Al-, Fe- and Mg-OH for phyllosilicates ^[12], Si-OH for hydrated silica ^[13], and C-O for carbonate ^[12]. Exceptions to this second rule are some Ca-sulfates, which do present a large triple absorption between 2.1 and 2.3 µm (S-O or OH/H₂O combinations/overtones) ^[11].

We begin by investigating the 2.4 µm drop slope of sulfates in contrast to the slope of their 2.2-2.3 µm plateau. To avoid contributions from absorption bands when computing the slopes, we use the upper part of a convex hull from 2.05 to 2.40 µm to represent the spectrum as if no absorptions features were present there. From this convex hull, we compute two slopes, one between 2.22 and 2.24 µm (“Plateau slope”, green line on Figure 1), and the other between 2.36 and 2.39 µm (“Drop slope”, blue line on Figure 1). We then subtract the drop slope to the plateau slope, obtaining our first parameter named “DIFFSLOPE”. We correlate this value to the depth of a potential absorption band in the 2.2-2.3 µm range, lacking for most sulfate species. For this, we divide our original spectrum by the same convex hull, and measure the depth of an eventual feature as our second parameter (“Band depth”, red dotted line on Figure 1).

These parameters are applied on lab spectra, and preliminary results are shown in Figure 2. Gypsum and anhydrite (blue circle) plot higher than other sulfates (green circle), due to their absorption in the 2.2-2.3 µm range. However, because of their strong drop slope, they still plot away from other hydrated phases. Overall, we obtain a clear separation between sulfates and other hydrated minerals, which is a promising first step before applying this method to Mars data.

We then plan to apply our method to the VISIR data from the SuperCam ^[14,15] instrument onboard the Perseverance rover. By combining improved VISIR spectral parameters and results from SuperCam’s other analytical techniques (mainly LIBS and Raman), we aim to revisit the in-situ detections of hydrated sulfates in Jezero crater ^[16], in order to better understand the aqueous history of the site.

Aknowledgments: This work is supported by the ERC OCEANID project (Grant agreement ID: 101045260) funded by the Horizon programme from the European Research Council.

References: ^[1]Bąbel and Schreiber (2014), Elsevier, ^[2]Murchie et al. (2009), J. Geophys. Res., ^[3]Carter et al. (2013), J. Geophys. Res., ^[4]Ehlmann and Edwards (2014), Annu. Rev. Earth Planet. Sci., ^[5]Milliken et al. (2009), Geophys. Res. Lett., ^[6]Cloutis et al. (2008), Icarus, ^[7]Poitras et al. (2018), Icarus, ^[8]Stack and Milliken (2015), Icarus, ^[9]Baschetti et al. (2026), Icarus, ^[10]Viviano-Beck et al. (2014), JGR Planets, ^[11]Cloutis et al. (2006), Icarus, ^[12]Clark et al. (1990), J. Geophys. Res., ^[13]Goryniuk et al. (2004), Geophys. Res. Lett, ^[14]Wiens et al. (2020), Space Sci. Rev., ^[15]Maurice et al. (2021), Space Sci. Rev., ^[16]Mandon et al. (2023), JGR Planets.

Figure 1 : Illustration of parameters computed in this study on example gypsum and epsomite lab spectra.

Figure 2 : Scatter plot representing the 2 parameters mentionned in the text against each other, for different mineralogical families.