Introduction
Clay minerals are ubiquitous across Earth and have also been detected on Mars (Ehlmann and Edwards, Reference Ehlmann and Edwards2014; Bishop et al., Reference Bishop, Fairén, Michalski, Gago-Duport, Baker, Gross, Velbel and Rampe2018; Tu et al., Reference Tu, Rampe, Bristow, Thorpe, Clark, Castle and Bedford2021b) and other planetary bodies (Rivkin et al., Reference Rivkin, Volquardsen and Clark2006; Ammannito et al., Reference Ammannito, DeSanctis, Ciarniello, Frigeri, Carrozzo, Combe and Raponi2016). Clay minerals are also used in many different applications such as environmental remediation and protection, agriculture, and manufacturing and as additives in drilling fluids and paper (Schoonheydt, Reference Schoonheydt2016). Within this study, the term ‘clay’ is used to refer to a clay mineral, e.g. smectite, rather than a particular grain size. Smectites are known to react with brines and there are many environments where a complex relationship exists between smectites and salts/brines, including evaporitic playa environments (Rosen, Reference Rosen1994) and cold aqueous systems such as Don Juan Pond in Antarctica and sediments on the surface of Mars (Schwenzer et al., Reference Schwenzer, Abramov, Allen, Bridges, Clifford, Filiberto and Newsom2012; Wilson et al., Reference Wilson, Wilson, Patey and Shaw2014; Rapin et al., Reference Rapin, Ehlmann, Dromart, Schieber, Thomas, Fischer and Clark2019; Tu et al., Reference Tu, Ming and Sletten2021a). In extremely cold environments, free liquid water is uncommon and increased relative humidity as a result of diurnal or seasonal variations becomes a significant factor when considering water–rock interactions (Polkko et al., Reference Polkko, Hieta, Harri, Tamppari, Martínez, Viúdez‐Moreiras and De La Torre Juarez2023). The hygroscopic nature of some salts leads to salt deliquescence during periods of sufficiently elevated relative humidity resulting in interactions between brines and smectite minerals (Brass, Reference Brass1980; Dickinson and Rosen, Reference Dickinson and Rosen2003; Gough et al., Reference Gough, Chevrier and Tolbert2014).
Although smectites, and associated cation exchange reactions, have been the focus of intense study for decades, new technology and techniques allow us to observe previously studied phenomena in new ways. One such application is environmental scanning electron microscopy (ESEM) which provides a unique opportunity to observe dynamic reactions in almost real time at micrometer to nanometer scales. Fundamentally, ESEM is an extension of scanning electron microscopy (SEM) in which a high level of vacuum is not required within the sample chamber (Danilatos, Reference Danilatos1988; Goldstein et al., Reference Goldstein, Newbury, Michael, Ritchie, Scott, Joy and Ritchie2018). In addition, the researcher can often control the environmental conditions experienced by the sample including precise control of water vapor pressure and temperature in the sample chamber. Therefore, ESEM samples do not experience desiccation as a result of high vacuum and the sample preparation requirements are substantially less complex compared with traditional SEM; in some cases no preparation is required (Baker et al., Reference Baker, Uwins and Mackinnon1993). Although clays are electrically insulating and typically would experience a charging effect inherent in electron microscopy applications, the presence of water vapor in ESEM applications reduces this charging effect, allowing greater resolution imaging of clays without coating. Accessories such as a Peltier cooling stage allow the researcher to modify experimental conditions in situ. By modifying conditions inside the ESEM chamber, dynamic experiments can be conducted while obtaining high-quality microscopic imaging of the reactants and products in situ, as the reaction is occurring.
Since the early 1990s, researchers have used ESEM to investigate the effects of relative humidity on smectite swelling, including studies focused on fluid–rock interactions within hydrocarbon reservoirs (Mehta, Reference Mehta1991; Uwins et al., Reference Uwins, Baker and Mackinnon1993; Baker et al., Reference Baker, Uwins and Mackinnon1994). Additional studies have examined the effects of clay–water vapor reactions, including efforts to determine the mechanism(s) responsible for damage to ancient Egyptian sculptures stored in a museum. By subjecting sepiolite and palygorskite clays to cyclical wetting and drying cycles, with and without NaCl salt present, Rodriguez-Navarro et al. (Reference Rodriguez-Navarro, Sebastian, Doehne and Ginell1998) confirmed visually that swelling at elevated relative humidities was amplified in the presence of NaCl salt.
Most clay studies using ESEM and varying relative humidity have overwhelmingly focused on clay swelling (Mehta, Reference Mehta1991; Uwins et al., Reference Uwins, Baker and Mackinnon1993; Baker et al., Reference Baker, Uwins and Mackinnon1994; Carrier et al., Reference Carrier, Wang, Vandamme, Pellenq, Bornert, Tanguy and Van Damme2013; Sun et al., Reference Sun, Mašín, Najser, Neděla and Navrátilová2019). However, in the presence of sorbed water or brine formed via salt deliquescence, cation exchange may also occur. Cation exchange is the process by which cations present in an aqueous solution undergo exchange with labile cations in a clay (Grim, Reference Grim1968). On Mars, diagenetic processes may have included acidic leaching, cation exchange, and limited mass transport due to low water:rock volumes (Vaniman et al., Reference Vaniman, Bish, Chipera and Rearick2011; Yen et al., Reference Yen, Ming, Vaniman, Gellert, Blake, Morris and Edgett2017; Geyer et al., Reference Geyer, Madden, Rodriguez, Bishop, Mason and Madden2023). The effects of cation exchange processes on the smectite and surrounding sediments largely depend on the specific conditions present. A series of previous studies examined relative humidity cycling of smectite-salt mixtures in order to better understand the evolution of CaSO4 and MgSO4 hydrates like those observed on Mars (Vaniman and Chipera, Reference Vaniman and Chipera2006; Vaniman et al., Reference Vaniman, Bish, Chipera and Rearick2011; Wilson and Bish, Reference Wilson and Bish2011; Vaniman et al., Reference Vaniman, Martínez, Rampe, Bristow, Blake, Yen and Morookian2018). The experiments used X-ray diffraction (XRD) to detect sulfate mineral changes as a result of varied relative humidity and found that relative humidity cycling not only affected sulfate mineralogy but that the presence of smectite also had an influence. Wilson and Bish (Reference Wilson and Bish2011) attributed the appearance of hydrated CaSO4 to cation exchange between deliquesced MgSO4 salts and Ca-bearing smectite in the absence of free liquid water. Although the results of the above-mentioned studies shed light on the relationship between relative humidity and cation exchange, XRD measurements did not allow direct observation of the mechanisms driving cation exchange and the spatial relationships of subsequent secondary minerals. A conceptual model of angstrom-thick briny films reacting with minerals at both Antarctica and on Mars has been proposed by Dickinson and Rosen (Reference Dickinson and Rosen2003). By studying ground-ice that forms as pore or fracture filling in soils at Dry Valleys, Antarctica, they demonstrated that the ground-ice was unlikely to be a product of meltwater, instead arguing that deliquescence of salts driven by relative humidity was likely the source of the ground-ice. While the study performed by Dickinson and Rosen is supported by bulk chemical analysis and isotopes of ground-ice, as well as static SEM mineral images, direct evidence linking relative humidity and reactions occurring in Antarctic soils is absent (Dickinson and Rosen, Reference Dickinson and Rosen2003).
The purpose of this study was to investigate the physiochemical effects of relative humidity on smectite-salt mixtures at the microscopic scale using ESEM. Understanding the effects of brine–smectite interactions is important for interpreting mineral assemblages and geomorphology formed in extremely cold aqueous environments, including Don Juan Pond and sedimentary systems on Mars. This study addresses the dearth of microscale, in situ observations of cation exchange reactions by examining the effects of relative humidity on smectite-salt mixtures via ESEM during temperature-driven humidity cycles.
Materials and methods
Experimental overview
In order to evaluate the effects of relative humidity (RH) on smectite particles in contact with salts and the potential for cation exchange, samples were observed using ESEM to collect high resolution secondary electron (SE) images and elemental maps of target smectite particles before, during, and after a 100% RH cycle. Preliminary experiments were conducted to determine optimal sample quantities and imaging parameters, while also defining equipment limitations. From preliminary experiments, 100 mg of sample was determined to be sufficient to prevent the formation of free liquid water within the duration of the experiments. In subsequent experiments, the sample quantity was increased to 2 g, and the time spent at 100% RH was reduced in order to prevent the formation of free liquid water.
Montmorillonite ‘SAz-1’ (Apache County, AZ, USA) from the Source Clays Repository of The Clay Mineral Society was used as the clay substrate in these experiments because it has been characterized extensively (Xu et al., Reference Xu, Johnston, Parker and Agnew2000; Chipera and Bish, Reference Chipera and Bish2001; Mermut and Cano, Reference Mermut and Cano2001; Mermut and Lagaly, Reference Mermut and Lagaly2001) and allows direct comparison with results of previous studies (Wilson and Bish, Reference Wilson and Bish2011; Reference Wilson and Bish2012). SAz-1 is a 2:1 phyllosilicate dioctahedral smectite (Na0.4(Al1.6Mg0.4)Si4O10(OH)2) with a relatively high cation exchange capacity (CEC) at 123 meq 100 g–1 (Borden and Giese, Reference Borden and Giese2001; Essington, Reference Essington2015). As with many 2:1 phyllosilicates, SAz-1 exhibits remarkable swelling when exposed to water, either as free liquid or vapor. This swelling is the result of water molecules sorbing between adjacent 2:1 layers; the water molecules may contribute to hydration of interlayer cations or be present as discrete layers of water in addition to the hydrated interlayer cations (Grim, Reference Grim1968; Moore and Reynolds, Reference Moore and Reynolds1997)
To create a consistent starting material, natural SAz-1 was soaked in saturated CaCl2 solution for >24 h to replace all exchangeable cations with Ca2+. The SAz-1 was then separated from the liquid by three rounds of centrifugation for 20 min at 10,000 rpm and rinsed with 18.2 MΩ-cm H2O, followed by freeze drying and gentle homogenization in a mortar and pestle. SAz-1 was mixed with ACS grade (>99.0% purity) Na2SO4 or MgSO4 anhydrous salts in a 1:10 salt:montmorillonite mass ratio. To form direct contacts between salt and montmorillonite grains as well as minimize air-filled pockets within the sample, a custom 3D-printed sample press was developed. This allowed uniform compression of the sample while conforming to the dimensions of the sample cup; by pressing the air out of the sample during sample preparation, minimal decompression occurred during the initial vacuum conditions inside the ESEM sample chamber. In addition, the sample press produced a relatively flat surface on the sample which made it significantly easier to image in the microscope. After the sample was mixed and pressed it was immediately loaded into the ESEM chamber.
Na2SO4 and MgSO4 salts were chosen because: (1) they have been observed on Mars (Hecht et al., Reference Hecht, Kounaves, Quinn, West, Young, Ming and Hoffman2009); and (2) the deliquescence points of both are well known and are within the range of relative humidity observed on present-day Mars. Therefore, they may have served an important role in Martian rheological and diagenetic processes (Brass, Reference Brass1980; Vaniman et al., Reference Vaniman, Bish, Chipera, Fialips, William Carey and Feldman2004; Möhlmann and Thomsen, Reference Möhlmann and Thomsen2011). Using two different salts also allowed us to compare the effects of cation chemistry on deliquescence and cation exchange behavior. Generally, for electrolyte solutions with low cation concentrations mixed with montmorillonite, cation selectivity preferences of montmorillonite follow Ca2+>Mg2+>Na+, where Ca2+ is retained in the montmorillonite interlayer most strongly and Na+ is most weakly retained in the montmorillonite (Sposito et al., Reference Sposito, Holtzclaw, Charlet, Jouany and Page1983a; Tang and Sparks, Reference Tang and Sparks1993; Appelo and Postma, Reference Appelo and Postma2004). However, when the electrolyte solution is not dilute, but instead the cation in solution is at such a high concentration as to be considered near-infinite, montmorillonite cation selectivity plays a less significant role. Instead, the dominant influence of which cation is adsorbed in the montmorillonite interlayer is controlled by the law of mass action. If, for example, a Ca2+-saturated montmorillonite was mixed with a near-saturated Na2SO4 solution, then some of the adsorbed Ca2+ cations would exchange with the Na+ cations in solution (Sposito et al., Reference Sposito, Holtzclaw, Jouany and Charlet1983b). Following the exchange, the Ca2+ now in solution could complex with the available SO42–, potentially leading to precipitation of CaSO4 minerals such as gypsum (CaSO4 • 2H2O), basanite (CaSO4 • ½H2O), and anhydrite (CaSO4).
Scanning electron microscopy
Imaging and analysis were accomplished using a ThermoFisher Scientific Quattro S field emission ESEM with a 20 kV accelerating voltage and a probe current of 4.1 mA; working distance varied depending on vacuum conditions. High-resolution SE images were collected at a chamber pressure of 70 Pa with a gaseous secondary electron detector operating in the secondary electron mode. Elemental analysis and mapping were accomplished using energy dispersive X-ray spectroscopy (EDS) in both high-vacuum and low-vacuum ESEM conditions. RH within the chamber was controlled by holding the chamber water vapor pressure constant at 800 Pa and cooling or heating the sample to a target temperature using a Peltier cold stage. The target temperature for a desired RH% was determined using the Antoine equation with updated parameters (NIST: https://webbook.nist.gov/cgi/cbook.cgi?ID=C7732185&Units=SI&Type=ANTOINE&Plot; Wood, Reference Wood1970); temperatures during the relative humidity cycles varied between 25.2°C and 3.8°C. Starting RH was ~3% and was increased to 100% RH over a period of 30 min, 100% RH was maintained for 1 h followed by a ramp back to ~3% RH over 30 min. Three per cent RH was chosen as the starting RH because this RH could be easily maintained in high-vacuum mode, allowing high-resolution imaging and X-ray analysis prior to and after exposure to 100% RH. In some cases, images of the montmorillonite particles are only presented after the RH ramp due to movement of the sample and loss of the original location.
Results
Initial montmorillonite characterization
The SAz-1 after Ca-saturation was fine grained with aggregates of montmorillonite particles averaging between 5 and 12 μm in diameter (Fig. 1). EDS elemental mapping prior to reaction confirmed that Ca was distributed homogeneously throughout the starting material. Magnesium was also abundant in SAz-1, likely in the octahedral sheet as noted elsewhere (Essington, Reference Essington2015; Mermut and Cano, Reference Mermut and Cano2001; Moore and Reynolds, Reference Moore and Reynolds1997) (Fig. 2).
Figure 1. SE image of the unaltered montmorillonite starting material.
Figure 2. Colorized EDS X-ray maps of the unaltered montmorillonite under initial conditions, from the same area as Fig. 1. Color intensity is proportional to the concentration of the indicated element at each pixel (Na and Fe also detected but not shown).
Preliminary experiment
A preliminary test was conducted to determine optimal sample conditions as well as identify the potential impact of equipment limitations. In the preliminary experiment ~100 mg of Na2SO4 salt+smectite was used to test the hypothesis that elevated RH within the instrument chamber would produce free bulk water in contact with the sample.
As expected, one of the most recognizable processes observed during the relative humidity cycling was swelling. In every experiment, as relative humidity increased, SAz-1 particles increased in diameter with a concurrent loss or reduction in sharpness of angular features; as relative humidity decreased, the same particles shrank in diameter and angular features regained their sharpness. Below 25% RH, no smectite volumetric changes were observed; any apparent sample shifting at <25% RH was attributed to thermal equilibration. Once RH increased beyond 25%, montmorillonite particles began to swell and shift, with individual montmorillonite particles increasing in diameter and roundness in real-time. While the montmorillonite particles equilibrated quickly (<30 s) at any given RH, the salt crystals took much longer to equilibrate. In many cases, it took >2 min for salt crystals to become visibly affected by the humidity, although the time was heavily dependent on the volume of the salt crystals.
Secondary electron images were collected throughout the hydration-dehydration cycles. In Fig. 3A, the sample was equilibrated at 25% RH and no bulk liquid water was observed. As the salt crystal equilibrated at 100% RH, roughening of the salt crystal surfaces became apparent (Fig. 3B), then as RH was held constant at 100%, a uniform layer of liquid formed, first on the Na2SO4 surfaces and eventually engulfed the montmorillonite particles (Fig. 3C). The bulbous, nodular shapes shown on the Na2SO4 crystal face in Fig. 3C were interpreted as a homogeneous water layer rather than a solid. As the RH increased, the nodules grew laterally and vertically like a film. Eventually, this nodular film grew to the point where it came into visible contact with the montmorillonite grains and began to cover them. Finally, as RH was brought back down to 25% the film covering the surfaces shrank in a similar manner to the growth stage, but in the reverse direction as RH decreased. The remaining bulk liquid water flash evaporated or co-precipitated at a point between 28% RH and 24% RH, leaving behind a network of nano- to microscale material that likely formed from widespread dissolution–precipitation (Fig. 3D).
Figure 3. SAz-1 on a Na2SO4 crystal with yellow arrows indicating the progression through one humidity cycle, starting at A and finishing at D. (A) 25% RH prior to exposure to elevated RH. (B) 100% RH after 30 min, observable montmorillonite swelling and increased surface roughness of the Na2SO4 crystal face. (C) 100% RH after 2.5 h, montmorillonite swelling had ceased, yet clear edges and particle shapes were maintained in the montmorillonite particles. No distinct features of the Na2SO4 crystal remained. Instead, the salt was covered with a bulbous, nodular film. (D) Return to 25% RH after 2.5 h at 100% RH, a network of nano- to micro-scale crystals, likely re-crystallized Na2SO4 appeared, no further changes were observed after return to 3% RH.
It is important to note the presence of a thick layer of bulbous material, which we interpret as bulk liquid water, visible in Fig. 3C. We hypothesize that the small sample volume and extended time at 100% RH in the preliminary experiment allowed a layer of bulk water to accumulate. However, in subsequent experiments no bulk water was observed, likely due to the larger sample volumes and reduced time spent at 100% RH. These subsequent experiments were performed using fresh, unreacted salt-montmorillonite mixtures.
Na2SO4 experiments in the absence of bulk water
A clear area with montmorillonite particles or isolated aggregates in contact with a Na2SO4 crystal were identified in order to study montmorillonite–salt interactions in situ (Fig. 4). The Na2SO4 crystal and montmorillonite composition was confirmed using EDS and detailed images collected before and after exposure to 100% RH (Target Grain 1: ‘TG1’; Figs 4–6). As humidity increased to 100% RH, significant lateral and vertical motion of the sample was observed, which was attributed to montmorillonite swelling due to increased hydration of the interlayer cations and/or additional interlayer water layers (Slade et al., Reference Slade, Quirk and Norrish1991). However, at no point during the relative humidity cycle was liquid water observed, even at 100% RH. Significant alteration of both the salt and montmorillonite were observed, however, most notably the formation of a salt dissolution pit which formed around a montmorillonite particle shown in Fig. 4B.
Figure 4. SE images of the target site consisting of montmorillonite clumps lying on top of a Na2SO4 crystal face, red boxes in panels A and B highlighting Target Grain 1 (TG1). (A) Initial conditions. (B) After exposure to 100% RH, severe dissolution etching and crystal formation were observed; the dashed green ovals highlight secondary crystals further detailed in Fig. 5.
After exposing the sample to 100% RH and then decreasing RH back to ~3%, new crystals appeared where montmorillonite particles were in contact with the Na2SO4 salt. It is worth noting that the new crystals only occurred where the montmorillonite was in direct contact with the salt; portions of the montmorillonite which were physically distant from the salt showed no signs of crystal nucleation or alteration other than residual swelling (Fig. 5). Large portions of the Na2SO4 crystal face also appeared to have been etched, particularly around TG1, but also around some of the other montmorillonite grains. One particularly notable dissolution pit surrounding TG1 had an indentation that mirrored the growth of the crystal attached to TG1 (just above the red rectangle in Fig. 4B). Additionally, the vertical topographic trench that crosses vertically through the field of view is visibly wider after 100% RH exposure.
Figure 5. Close-up of new crystals depicted in Fig. 4B, TG1 present at top right of both images. (A) SE image, green dashed ovals indicate new crystals. (B) Color EDS image of Ca for the same area as Fig. 4A. Note the presence of SAz-1 grains visible in the SE image but undetectable by Ca EDS, indicating that Ca has migrated out of the SAz-1 grains and into the surrounding secondary salt crystals precipitated on the surface of the montmorillonite.
In addition to evidence of etching and dissolution, new crystals also appeared after exposure to 100% RH. These new crystals exhibited completely different crystal habit relative to the Na2SO4 salt and also had a different chemical composition. EDS elemental mapping (Fig. 6) showed before and after element maps of TG1. Prior to the 100% RH ramp, Ca was distributed homogeneously in TG1; however, after reaching 100% RH, the Ca appeared to migrate from the center of the montmorillonite to the new crystals which formed at the contact between the montmorillonite and the Na2SO4 crystal. Unfortunately, crystal nucleation and growth occurred too rapidly to capture images of growth dynamics.
Figure 6. Ca migration observed by comparing before and after color EDS mapping of TG1. The top row images were collected prior to 100% RH, the bottom row images were collected after the sample was exposed to 100% RH and returned to ~3% RH.
SAz-1 grains that were isolated from the Na2SO4 failed to produce new crystals and the Ca distribution remained homogeneous throughout the experiment (Fig. 7). However, these isolated montmorillonite clumps did exhibit signs of swelling.
Figure 7. An isolated montmorillonite grain located on the aluminum sample holder wall away from any Na2SO4 salt. SE image (left) and color EDS – Ca image (right) collected after the grain was exposed to 100% RH. Note the homogeneous distribution of Ca in the sample.
MgSO4+SAz-1 interactions
When the montmorillonite was mixed with MgSO4, changes similar to those observed in the SAz-1+Na2SO4 experiments occurred. However, in the MgSO4 experiments, greater movement of the montmorillonite–salt mixture occurred as the relative humidity decreased. This prevented the collection of images and elemental maps from being compared directly with the same particle with initial conditions. Luckily, montmorillonite grains that had been in contact with salt crystals during the RH cycle were ubiquitous, allowing interpretation of reactions which occurred as a result of exposure to elevated relative humidity.
Similar to the Na2SO4 experiments, EDS elemental maps and SE images of TG2+MgSO4 were collected before and after RH cycling. Like the Na2SO4 experiments, blade-like crystals that contained high concentrations of Ca and S, but did not contain abundant Al or Si, appeared after the RH cycle, indicating that the newly formed Ca-rich crystals were related to, but not part of, the SAz-1 particle (Figs 8 and 9).
Figure 8. Elemental distributions in SAz-1+MgSO4 after one RH cycle. The red dashed boxes highlight crystals that exhibited co-localized Ca and S, but were lacking in Al or Si. The yellow dashed box in the SE image highlights TG2 shown in Fig. 9.
Figure 9. Close-up of TG2 showing differences in elemental distribution; yellow arrows indicate the location of crystals on the upper left of the particle, visible in SE, which contain abundant Ca and S based on the EDS maps, but are absent in the Si EDS map. The yellow dashed polygon outlines the montmorillonite grain as defined by the Si EDS map.
Discussion
Cation exchange and elemental migration
In both the Na2SO4 and MgSO4 experiments, Ca appeared to be concentrated in new crystals observed adjacent to montmorillonite grains (Figs 5 and 9), but only when the montmorillonite grains were in contact with salt crystals during a RH cycle. Ca was homogeneously distributed within montmorillonite grains prior to the relative humidity cycles with salts present (Fig. 2), while after the cycle Ca was no longer homogeneously distributed within the montmorillonite grain (Figs 6 and 9). This migration of Ca coincides with a co-localized concentration of sulfur (Figs 8 and 9), likely due to growth of CaSO4 crystals at the interface between the salt and the montmorillonite grains. Similar CaSO4 precipitation has been previously observed in XRD experiments investigating smectite-salt interactions during relative humidity changes (Wilson and Bish, Reference Wilson and Bish2011; Wilson and Bish, Reference Wilson and Bish2012). Based on this co-localization of S and Ca, the Na2SO4 and MgSO4 salts likely deliquesced forming a sub-micron layer of brine that facilitated cation exchange between the Ca2+ present in the montmorillonite interlayers and the cations in the salts. The exchange of Ca2+ from the montmorillonite interlayers resulted in saturation and precipitation of CaSO4, while Na+ or Mg2+ would have migrated from the salt into the montmorillonite to maintain the montmorillonite layer charge neutrality. Element migration with no liquid water observable, even at the sub-micrometer scale, also corroborates the work of Wilson and Bish (Reference Wilson and Bish2011) that indicated cation exchange can occur even within a molecular-scale surficial water layer.
Applications on Mars
Recurring slope lineae (RSL) observed on Mars, as well as water tracks seen at Don Juan Pond, may not only be indicative of fluid transport but also chemical alteration, albeit on a microscopic scale. While hypotheses differ regarding the cause(s) of RSL, such as if liquid water is present, one explanation is the deliquescence of salts during periods of high relative humidity (Dickson et al., Reference Dickson, Head, Levy and Marchant2013; Gough et al., Reference Gough, Wong, Dickson, Levy, Head, Marchant and Tolbert2017). Studies examining the water source for Don Juan Pond concluded that flows similar to RSL observed on Mars are likely due to the relative humidity controlled deliquescence of salts (Dickson et al., Reference Dickson, Head, Levy and Marchant2013; Gough et al., Reference Gough, Wong, Dickson, Levy, Head, Marchant and Tolbert2017). Tu et al. (Reference Tu, Ming and Sletten2021a) found that the presence of elevated Ca2+ concentrations in brines at Don Juan Pond is also likely due to cation exchange. When exposed to solutions with sufficient Na+ concentrations, some smectites will release Ca2+ in exchange for Na+ (McBride, Reference McBride1980; Toner and Sletten, Reference Toner and Sletten2013; Tu et al., Reference Tu, Ming and Sletten2021a), even though the smectite present at that location would theoretically retain Ca2+ preferentially over Na+.
A common contention to the briny RSL origin hypothesis is that RSL do not correlate perfectly with conditions conducive to the presence of liquid water (see McEwen et al., Reference McEwen, Ojha, Dundas, Mattson, Byrne, Wray and Gulick2011; Toner et al., Reference Toner, Sletten, Liu, Catling, Ming, Mushkin and Lin2022, and references within both for a review). Toner et al. (Reference Toner, Sletten, Liu, Catling, Ming, Mushkin and Lin2022) argued against cation exchange as the mechanism for RSL at both Don Juan Pond and on Mars due to the preferential adsorption of Ca2+ by smectites when exposed to dilute solutions containing the cations in question. However, as demonstrated by the result in the present study, when brine formation is initiated by deliquescence of salts controlled by relative humidity, Ca can migrate out of the smectite interlayer via cation exchange and even mass transport is possible, albeit on a microscopic scale.
The occurrence of cation exchange outside typical boundary conditions that assume bulk liquid water must be present brings with it the realization that Martian sediments may in fact be undergoing active diagenesis at the microscopic scale, even under traditionally ‘dry’ conditions. For example, hydrated salts have been detected at several RSL locations on Mars (Ojha et al., Reference Ojha, Wilhelm, Murchie, McEwen, Wray, Hanley and Chojnacki2015). Heinz et al. (Reference Heinz, Schulze‐Makuch and Kounaves2016) observed that a Mars analogue soil, when exposed to elevated RH, would only show darkening similar to RSL if salts were present and permitted to deliquesce. Thus, observations of RSL on Mars suggest chemical alteration may be occurring on diurnal and seasonal time scales. Similarly, cation exchange in high-salinity fluids may also have contributed to large CaSO4 concentrations observed across Mars soils, including CaSO4-rich veins (McLennan et al., Reference McLennan, Anderson, Bell, Bridges, Calef, Campbell and Cousin2014; Rampe et al., Reference Rampe, Blake, Bristow, Ming, Vaniman, Morris and Tu2020). The elemental composition of sediments at Gale crater suggests limited chemical transport and low water:rock volumes (Bristow et al., Reference Bristow, Bish, Vaniman, Morris, Blake, Grotzinger and Ming2015); these observations are consistent with the hypothesis that cation exchange played a major role in the formation of the CaSO4 minerals.
Conclusions
ESEM observations revealed how minerals are altered or formed as a direct result of exposure to elevated RH. The results of the present study include a before and after set of SE and elemental images which document the dissolution of salt in the absence of free liquid water, which corresponds with elemental migration consistent with cation exchange between a brine and montmorillonite grains. New crystal formation observed in SE images was further distinguished by collocated concentrations of Ca and sulfur observed in elemental EDS maps, suggesting the formation of a CaSO4-rich phase. In the case of both Na2SO4 and MgSO4 salts, after a single RH cycle, montmorillonite grains that had been saturated with Ca and were in direct physical contact with salt grains no longer exhibited a homogeneous Ca distribution. Instead, Ca appeared to migrate to the edge of the grain and into the newly formed crystals. The results of this study thus provide evidence that sub-micrometer thick layers of brine formed by salt deliquescence can facilitate cation exchange and subsequent mineral formation in extreme environments such as Antarctica or Mars.
Author contributions
Christopher Geyer conceived, planned, and executed the experiments under the guidance of Megan Elwood Madden, Andrew S. Elwood Madden, and Preston R. Larson. Preston R. Larson guided instrument operation and image acquisition. Megan Elwood Madden and Andrew S. Elwood Madden supervised data interpretation. All authors provided quality feedback and significantly influenced the research and manuscript.
Acknowledgements
We thank the editors of Clays and Clay Minerals as well as three anonymous reviewers for feedback that improved the manuscript.
Financial support
This project was funded by the University of Oklahoma and NASA Solar System Workings Grant 80NSSC23K0037.
Competing interests
The authors declare that they have no competing interests.
Data availability statement
Images used in this research are available upon request
Open access
