Introduction
Phyllosilicates are an important mineral group in ancient (~3.5–4.1 Ga, based on impact crater counting) Martian strata because they are indicative of water–rock interactions and can help constrain aqueous conditions that may have been habitable by microbial life. Orbital visible/shortwave-infrared (VSWIR) reflectance data collected by the Observatoire pour la Mineralogie, l’Eau, les Glaces et l’Activité (OMEGA), and the Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) indicate that smectite is the most common phyllosilicate group on the Martian surface, and Fe/Mg-smectite (e.g. nontronite and saponite) is much more common than Al-smectite (e.g. montmorillonite) (e.g. Poulet et al., Reference Poulet, Bibring, Mustard, Gendrin, Mangold, Langevin, Arvidson, Gondet and Gomez2005; Mustard et al., Reference Mustard, Murchie, Pelkey, Ehlmann, Milliken, Grant, Bibring, Poulet, Bishop, Noe Dobrea, Roach, Seelos, Arvidson, Wiseman, Green, Hash, Humm, Malaret, McGovern, Seelos, Clancy, Clark, Des Marais, Izenberg, Knudson, Langevin, Martin, McGuire, Morris, Robinson, Roush, Smith, Swayze, Taylor, Titus and Wolff2008; Ehlmann et al., Reference Ehlmann, Mustard, Murchie, Bibring, Meunier, Fraeman and Langevin2011). The types and diversity of phyllosilicates, however, depend on the geologic environment, where phyllosilicates in igneous settings are dominated by Fe/Mg-smectite and chlorite, and phyllosilicates in sedimentary rocks and stratigraphic sections generally contain a mix of Fe/Mg-smectite and Al-phyllosilicates (e.g. montmorillonite and kaolinite) (Ehlmann et al., Reference Ehlmann, Mustard, Murchie, Bibring, Meunier, Fraeman and Langevin2011; Ehlmann et al., Reference Ehlmann, Berger, Mangold, Michalski, Catling, Ruff, Chassefière, Niles, Chevrier and Poulet2013).
Smectite has been identified in abundances of up to ~30 wt.% in early Hesperian-aged (~3.6 Ga old) lacustrine mudstone in Gale crater by the CheMin X-ray diffractometer on the Mars Science Laboratory Curiosity rover (e.g. Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014; Bristow et al., Reference Bristow, Rampe, Achilles, Blake, Chipera, Craig, Crisp, Des Marais, Downs, Gellert, Grotzinger, Gupta, Hazen, Horgan, Hogancamp, Mangold, Mahaffy, McAdam, Ming, Morookian, Morris, Morrison, Treiman, Vaniman, Vasavada and Yen2018; Rampe et al., Reference Rampe, Blake, Bristow, Ming, Vaniman, Morris, Achilles, Chipera, Morrison, Tu, Yen, Castle, Downs, Downs, Grotzinger, Hazen, Treiman, Peretyazhko, Des, Marais, Walroth, Craig, Cirsp, Lafuente, Morookian, Sarrazin, Thorpe, Bridges, Edgar, Fedo, Freissinet, Gellert, Mahaffy, Newsom, Johnson, Kah, Siebach, Schieber, Sun, Vasavada, Wellington and Wiens2020a; Tu et al., Reference Tu, Rampe, Bristow, Thorpe, Clark and Castle2021; Thorpe et al., Reference Thorpe, Bristow, Rampe, Tosca, Grotzinger, Bennett, Achilles, Blake, Chipera, Downs, Downs, Morrison, Tu, Castle, Craig, Des, Marais, Hazen, Ming, Morris, Treiman, Vaniman, Yen, Vasavada, Dehouck, Bridges, Berger, McAdam, Peretyazhko, Siebach, Byrk, Fox and Fedo2022). The composition and structure of the smectite identified in Gale crater changes throughout a 400+ m vertical stratigraphic succession of fluvial-lacustrine deposits. Trioctahedral Fe3+-bearing saponite, from oxidation of Fe2+ in the octahedral site, is present at the base of the section (Treiman et al., Reference Treiman, Morris, Agresti, Graff, Achilles, Rampe, Bristow, Ming, Blake, Vaniman, Bish, Chipera, Morrison and Downs2014). Mixed dioctahedral/trioctahedral smectite is present in the middle of the section, and fully dioctahedral nontronite is present in the Glen Torridon region (Bristow et al., Reference Bristow, Rampe, Achilles, Blake, Chipera, Craig, Crisp, Des Marais, Downs, Gellert, Grotzinger, Gupta, Hazen, Horgan, Hogancamp, Mangold, Mahaffy, McAdam, Ming, Morookian, Morris, Morrison, Treiman, Vaniman, Vasavada and Yen2018; Bristow et al., Reference Bristow, Grotzinger, Rampe, Cuadros, Chipera, Downs, Fedo, Frydenvang, McAdam, Achilles, Blake, Castle, Craig, Des Marais, Downs, Hazen, Ming, Morris, Morrison, Thorpe, Treiman, Tu, Vaniman, Yen, Gellert, Mahaffy, Wiens, Bryk, Bennett, Fox, Milliken, Fraeman and Vasavada2021; Tu et al., Reference Tu, Rampe, Bristow, Thorpe, Clark and Castle2021; Thorpe et al., Reference Thorpe, Bristow, Rampe, Tosca, Grotzinger, Bennett, Achilles, Blake, Chipera, Downs, Downs, Morrison, Tu, Castle, Craig, Des, Marais, Hazen, Ming, Morris, Treiman, Vaniman, Yen, Vasavada, Dehouck, Bridges, Berger, McAdam, Peretyazhko, Siebach, Byrk, Fox and Fedo2022). All phyllosilicates measured by CheMin have a broad basal peak at ~10 Å and relatively low K2O content, indicating the presence of collapsed smectite as opposed to illite (e.g. Bristow et al., Reference Bristow, Rampe, Achilles, Blake, Chipera, Craig, Crisp, Des Marais, Downs, Gellert, Grotzinger, Gupta, Hazen, Horgan, Hogancamp, Mangold, Mahaffy, McAdam, Ming, Morookian, Morris, Morrison, Treiman, Vaniman, Vasavada and Yen2018), except for one sample at the base of the stratigraphic section. This sample, named ‘Cumberland’, was drilled from a formation called Yellowknife Bay and has a clearly defined basal spacing of 13.5 Å (Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014) (Fig. 1). This peak position remained constant for >150 Martian days in the warm and dry environment within the CheMin instrument, indicating that the phyllosilicate did not collapse from dehydration over time.
Figure 1. Yellowknife Bay drill site and CheMin data. (A) Mastcam mosaic of Yellowknife Bay with the geologic members and drill sites identified. Image credit: NASA/JPL/MSSS. (B) CheMin X-ray diffraction patterns of the John Klein and Cumberland drill targets. (C) Portion of John Klein and Cumberland CheMin XRD patterns showing the positions of (001) and (02l) peaks.
Constraining the composition and structure of phyllosilicates on the Martian surface permits the characterization of aqueous environments in which they formed and development of a more complete geologic history of the planet. Fe/Mg-smectite commonly forms from mafic parent material, which is abundant on Mars. The abundant Fe/Mg-smectites on Mars probably formed from a variety of processes, including magmatic processes (e.g. deuteric alteration), hydrothermalism, metamorphism, precipitation in lake sediments, surficial weathering profiles, and, perhaps, under a steamy primordial atmosphere (e.g. Ehlmann et al., Reference Ehlmann, Berger, Mangold, Michalski, Catling, Ruff, Chassefière, Niles, Chevrier and Poulet2013; Cannon et al., Reference Cannon, Parman and Mustard2017; Mangold et al., Reference Mangold, Dehouck, Fedo, Forni, Achilles and Bristow2019). Al-phyllosilicates, like montmorillonite and kaolinite, are more common in pedogenic and open-system weathering environments where leaching removes basic cations (e.g. Meunier, Reference Meunier2005; Ehlmann et al., Reference Ehlmann, Berger, Mangold, Michalski, Catling, Ruff, Chassefière, Niles, Chevrier and Poulet2013). Chlorite is indicative of diagenesis or metamorphism, but chlorite can also form in pedogenic environments (e.g. Barnhisel and Bertsch, Reference Barnhisel, Bertsch, Dixon and Weed1989; Meunier, Reference Meunier2005).
Smectite can be partially or fully chloritized in soil environments on Earth. Partially chloritized smectite (also known as hydroxy-interlayered smectite) typically forms through the replacement of solvated cations in smectite interlayer sites by hydroxy-Al, -Mg, or -Fe polymeric components (Barnhisel and Bertsch, Reference Barnhisel, Bertsch, Dixon and Weed1989). Hydroxy-Al smectite is more common on Earth than hydroxy-Fe or hydroxy-Mg smectite because Al is more abundant in the Earth’s crust than Fe and Mg. It forms under moderately acidic pH, low organic content, frequent wetting and drying cycles, and oxidizing conditions. Mg- and Fe-partially chloritized smectite have been identified on Earth in alkaline, saline lake sediments (Jones and Weir, Reference Jones and Weir1983) and reducing sedimentary environments (Lynn and Whittig, Reference Lynn and Whittig1966), respectively. Partially chloritized smectite is structurally similar to mixed-layer chlorite-smectite. Mixed-layer chlorite-smectite is composed of complete chlorite and smectite layers stacked on top of one another in the c crystallographic direction. Partially chloritized smectite is smectite with incomplete gibbsite- or brucite-like sheets (or ‘pillars’ as they are sometimes referred to in the literature) in the interlayer site (Fig. 2), making partially chloritized smectite a type of incipient mixed-layer chlorite smectite. This interlayer structure allows the smectite to swell when treated with ethylene glycol but prevents it from fully collapsing when heated to 500°C (Barnhisel and Bertsch, Reference Barnhisel, Bertsch, Dixon and Weed1989).
Figure 2. Structural diagram of smectite and partially chloritized smectite. Left: diagram of saponite, a trioctahedral smectite, showing water molecules and solvated cations in the interlayer site; right: diagram of partially chloritized saponite, showing water molecules, solvated cations, and short domains of brucite sheets in the interlayer site.
Bristow et al. (Reference Bristow, Bish, Vaniman, Morris, Blake, Grotzinger, Rampe, Crisp, Achilles, Ming, Ehlmann, King, Bridges, Eigenbrode, Sumner, Chipera, Morookian, Treiman, Morrison, Downs, Farmer, Des Marais, Sarrazin, Floyd, Mischna and McAdam2015) hypothesized that the expanded smectite in Yellowknife Bay could be explained by partial chloritization by hydroxy-Mg. This idea is supported by the identification of Mg-rich diagenetic features in Yellowknife Bay (Léveillé et al., Reference Léveillé, Bridges, Wiens, Mangold, Cousin and Lanza2014). Furthermore, mafic silicate minerals, olivine and pyroxene, are abundant across the Martian surface because Mars is dominated by basalt and sediments derived from basalt (e.g. Rogers and Christensen, Reference Rogers and Christensen2007). As such, aqueous alteration on Mars results in more Mg2+, Fe2+, and Fe3+ in solution than is typically seen in continental terrestrial settings (e.g. Hurowitz and McLennan, Reference Hurowitz and McLennan2006). The Mars science community has previously lacked the data necessary to identify partially chloritized smectite on the Martian surface. The objective of this work was to further interpret the clay mineralogy of Gale crater to learn more about past aqueous environments. This work tested the hypothesis that expanded smectite in Yellowknife Bay is a result of partial chloritization by hydroxy-Mg, by measuring partially chloritized and Mg-saturated smectite using laboratory instruments that are analogous to those on Mars rovers and orbiters. Laboratory data are compared with data collected by instruments on the Curiosity rover to look for signatures of partial chloritization in Martian data.
Methods
Partial chloritization procedure
The samples studied were nontronite (sample NAu-2 from the Source Clay Repository of The Clay Minerals Society), montmorillonite (Source Clay SWy-1), and Fe-saponite (a Mars-analog smectite sourced from Griffith Park, California (Treiman et al., Reference Treiman, Morris, Agresti, Graff, Achilles, Rampe, Bristow, Ming, Blake, Vaniman, Bish, Chipera, Morrison and Downs2014), abbreviated to ‘GP’ here). The samples were selected as starting materials for partial chloritization experiments. The <2 μm size fraction of all samples was separated via centrifugation. Na-saturation (Moore and Reynolds, Reference Moore and Reynolds1997) was performed prior to partial chloritization to start with a homogeneous interlayer composition. Each smectite was Mg-saturated (Moore and Reynolds, Reference Moore and Reynolds1997) to provide a comparison to data from chloritized smectite. Mg-saturation was performed by distributing clay in a 0.1 M MgCl2 solution in an ultrasonicator for 2 min, and then shaking for 24 h. Clay suspensions were centrifuged, replenished with 0.1 M MgCl2 twice more, and then washed with deionized water and a 50:50 by volume mixture of deionized water and ethanol. Mg-saturated smectites were allowed to dry at 60°C for 24 h (Moore and Reynolds, Reference Moore and Reynolds1997).
Portions of the Mg-saturated smectite samples were chloritized with Mg(OH)2 as the ‘pillars’ in the interlayer site at various OH:Mg molar ratios to achieve various levels of chloritization. Molar OH:Mg ratios of 0.5, 1.0, 1.5, and 1.75 in the solution were chosen to mimic previous partial chloritization methods (Xeidakis, Reference Xeidakis1996). Increasing the OH:Mg ratio causes more Mg(OH)2 pillars to precipitate in the interlayer site, resulting in a greater degree of chloritization. Briefly, a clay suspension was made by adding ~0.8 g Mg-saturated smectite in 50 mL milli-Q water, and 1.5 M NaOH was added dropwise to desired OH:Mg molar ratios of 0.5, 1.0, 1.5, or 1.75. The clay suspension was stirred vigorously during titration, and the final product was centrifuged, washed twice with deionized H2O to remove soluble salts, and dried at 25°C.
X-ray diffraction
Random powder mounts of Mg-saturated and Mg(OH)2 chloritized smectite samples were analyzed at ~1% (in dry N2(g)) and ~90% relative humidity (RH) on a non-ambient Anton Paar stage in a Panalytical X’Pert Pro MPD X-ray diffractometer at the NASA Johnson Space Center (JSC) from 2 to 80°2θ (CoKα), with a 0.02°2θ step size, and 10 s step–1. The samples were held at the desired RH for at least 2 h before analysis. Random powder mounts, rather than oriented mounts, were measured because these are most analogous to the patterns measured by the CheMin instrument. Mg-saturated and chloritized samples were measured on a CheMin IV XRD instrument, a laboratory analog to the CheMin instrument on board Curiosity. Mg-saturated nontronite was measured using an Olympus Terra instrument because the CheMin IV was not in operation. Samples were heated under desiccating conditions in N2(g) at 200°C (i.e. ‘heat desiccated’) prior to analysis to drive off interlayer water and emulate the low hydration state of smectite on the Martian surface. CheMin IV measurements were made with dry N2(g) flowing over the samples. Unlike most laboratory X-ray diffractometers, CheMin IV operates in transmission mode, and data are collected from 4 to 55°2θ (Co-Kα) with a 0.05° step size (e.g. Blake et al., Reference Blake, Vaniman, Achilles, Anderson, Bish, Bristow, Chen, Chipera, Crisp, Des Marais, Downs, Farmer, Feldman, Fonda, Gailhanou, Ma, Ming, Morris, Sarrazin, Stolper, Treiman and Yen2012). CheMin uses a cobalt X-ray source to minimize fluorescence from Fe. Terra is a field-portable version of CheMin with the similar X-ray source, angular range and resolution, geometry, and sample handling system as CheMin (e.g. Sarrazin et al., Reference Sarrazin, Brunner, Blake, Gailhanou, Bish and Vaniman2008). Peak positions were calculated as the centroid position from Lorentz-polarization (LP)-corrected patterns using the Materials Data Inc. JADE program.
Thermal and evolved gas analysis
Mg-saturated and chloritized smectite samples were analyzed individually using a Setaram Labsys EVO thermal gravimeter (TG)/differential scanning calorimeter (DSC)/furnace connected to a Pfeiffer ThermoStar quadrupole mass spectrometer (QMS) at JSC configured to operate similarly to the Sample Analysis at Mars-Evolved Gas Analysis (SAM-EGA) instrument on board the Curiosity rover (e.g. Mahaffy et al., Reference Mahaffy, Webster, Cabane, Conrad, Coll, Atreya, Arvey, Bariciniak, Benna, Bleacher, Brinckerhoff, Eigenbrode, Carignan, Cascia, Chalmers, Dworkin, Errigo, Everson, Franz, Farley, Feng, Frazier, Freissinet, Glavin, Harpold, Hawk, Holmes, Johnson, Jones, Jordan, Kellogg, Lewis, Lyness, Malespin, Martin, Maurer, McAdam, McLennan, Nolan, Noriega, Pavlov, Prats, Raaen, Sheinman, Sheppard, Smith, Stern, Tan, Trainer, Ming, Morris, Jones, Gundersen, Steele, Wray, Botta, Leshin, Owen, Battel, Jakosky, Manning, Squyres, Navarro-Gonzalez, McKay, Paulin, Sternberg, Buch, Sorensen, Kline-Schoder, Soscia, Szopa, Teinturier, Baffes, Feldman, Flesch, Forouhar, Garcia, Keymeulen, Woodward, Block, Arnett, Miller, Edmonson, Gorevan and Mumm2012). EGA measures and quantifies the gases evolved during pyrolysis of a sample. TG/DSC data show sample weight loss and exothermic/endothermic reactions, respectively, during sample heating. Although the SAM instrument does not have TG/DSC capabilities, they are used in this study to better understand chemical reactions and phase transitions in the Mg-saturated and chloritized smectite. Samples of ~10 mg were placed in alumina ceramic crucibles. Sample and identical empty crucibles were placed inside the furnace, which was then purged with He gas and set to a pressure of 30 mbar and a flow rate of 10 standard cubic centimeters per minute (sccm). The crucibles were heated from ~30 to 1000°C at a heating rate of 35°C min–1. All samples were run in duplicate. EGA data from mass-to-charge ratios (m/z) of 2–100 were collected and mass-normalized, but only the results from m/z=18 (i.e. H2O) are presented here because it is specifically the water-release data that can be used to identify specific phyllosilicates (e.g. McAdam et al., Reference McAdam, Sutter, Archer, Franz, Wong, Lewis, Eigenbrode, Stern, Knudson, Clark, Andrejkovicova Slavka Ming, Morris, Achilles, Rampe, Bristow, Navarro-Gonzalez, Mahaffy, Thompson, Gellert, Williams, House and Johnson2020).
Visible/shortwave infrared (VSWIR) reflectance spectroscopy
VSWIR reflectance spectra (0.35–2.5 μm) were acquired at JSC on Mg-saturated and chloritized smectite samples with Analytical Spectral Devices (ASD) FieldSpec3 Hi-Res fiber optic spectrometers configured with Mug lights. Samples were measured under ambient laboratory conditions, after desiccation at 25°C in a glove box purged with dry N2(g) (~150–400 parts per million by volume (ppmv), H2O, or 0.5–2% RH), and after heating to 200°C under dry N2(g). Samples were desiccated in the dry N2(g) glove box at 25°C and were heated to 200°C under N2(g). Samples were held at each temperature until the spectra showed no changes (22–96 h for samples at 25°C and 24–384 h for samples at 200°C). For many Mg-saturated samples, there was insufficient material to exclude contributions from the sample substrate from the measured reflectance. The substrate material (LEE Filters Black Foil 280) has no absorptions relevant to the analysis but is dark, ~4% reflectance over the full range of the ASD instrument.
Results
This section describes the XRD patterns, EGA traces, and VSWIR spectra of the Mg-saturated and chloritized smectite samples. The results are separated by smectite type (i.e. Griffith Park saponite, NAu-2 nontronite, and SWy-1 montmorillonite). The results focus on: (1) the position of the (001) peaks in XRD patterns for comparison with the (001) peak in the Cumberland sample drilled on Mars; (2) the temperatures at which H2O was released in EGA for comparison with SAM EGA traces in Cumberland; and (3) changes in VSWIR spectra with increasing degree of chloritization.
X-ray diffraction
For all three smectites, partial chloritization led to changes to the (001) peak positions and no changes to the (02l) and (06l) peaks (Table 1; see also Table S1 in the Supplementary material), demonstrating that the chloritization procedure did not disrupt the b-axis of the smectite crystal structure. For each set of patterns measured under different conditions, a broad peak at ~7.4 Å appeared when smectite standards were subject to chloritization. The appearance of the (002) peak is a result of changes to the layer structure factor of smectites (e.g. Moore and Reynolds, Reference Moore and Reynolds1997) as Mg(OH)2 layers replace exchangeable interlayer sites. As described in more detail below, an increase in OH:Mg ratio in the chloritization procedure generally caused an increase in (002) height intensity to the (001) intensity (i.e. a decrease in the (001)/(002) intensity ratio), indicating a greater degree of chloritization with increase in OH:Mg ratio. The measured positions of the (001) peaks typically showed that complete chloritization was achieved with OH:Mg=1.5 and 1.75.
Table 1. LP-corrected d 001 and d 002 for Mg-saturated and chloritized smectite samples measured on the Panalytical at 90% RH and ~1% RH and heat-desiccated samples measured on the CheMin IV under dry N2(g)
Peak positions were measured using the centroid. Values are reported in Å. *NAu-2 measurements were made on an Olympus Terra instrument.
Griffith Park saponite
In all XRD measurements, the GP saponite has a (02l) peak at ~4.60 Å and a (06l) peak position of ~1.53 Å (Fig. 3; see also Table S1 and Fig. S3 in the Supplementary material), consistent with trioctahedral smectite (e.g. Brindley, Reference Brindley1952; Moore and Reynolds, Reference Moore and Reynolds1997; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014). For samples measured at 90% RH on the Panalytical XRD, the Mg-saturated sample was the most expanded with a position of 15.4 Å, whereas the chloritized samples had d 001≈14.3–15.0 Å (Fig. 3A). When measured under N2(g) on the Panalytical, the Mg-saturated and chloritized samples had similar d 001 values, from ~13.8 to 14.1 Å (Fig. 3B). Heat-desiccated samples measured under N2(g) on the CheMin IV showed that chloritized samples with greater OH:Mg ratios were the most expanded with d 001≈13.6–13.8 Å, the Mg-saturated sample is the least expanded with a d 001 of 9.9 Å, and the least chloritized sample (OH:Mg=0.5) has an intermediate d 001 of 12.2 Å (Fig. 3C). A weak, broad peak at ~4.9 Å is apparent in chloritized samples with OH:Mg=1.0, 1.5, and 1.75 in all patterns measured on the Panalytical and CheMin IV. The position of this peak is consistent with a (003) reflection.
Figure 3. XRD patterns from 5 to 30°2θ (20.5 to 3.5 Å) of Mg-saturated and chloritized GP saponite. LP-corrected d 001 of Mg-saturated samples and d 001, d 002, and d 02l of OH:Mg=1.75 samples are labeled for reference. (A) Patterns measured on the Panalytical at 90% RH. (B) Patterns measured on the Panalytical at ~1% RH (i.e. under dry N2(g)). (C) Patterns of heat-treated samples measured on the CheMin IV under dry N2(g).
NAu-2 nontronite
All XRD patterns from NAu-2 have a (02l) peak at ~4.53 Å and a (06l) peak at ~1.51 Å (Fig. 4; see also Table S1 and Fig. S2 in the Supplementary material), which are diagnostic of dioctahedral smectite (e.g. Brindley, Reference Brindley1952; Moore and Reynolds, Reference Moore and Reynolds1997; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014). The Mg-saturated NAu-2 nontronite sample measured at 90% RH on the Panalytical was the most expanded with a (001) peak position of 15.4 Å, and the chloritized nontronite samples have smaller d 001 at 13.7–14.5 Å (Fig. 4A). When measured under dry N2(g) on the Panalytical, the Mg-saturated NAu-2 sample has a very broad (001) peak with a maximum at 13.9 Å, suggesting a range of hydration states. The chloritized nontronite samples measured under N2(g) on the Panalytical show an increase in d 001 with the increase in extent of chloritization, from 13.9 Å for OH:Mg=0.5 to 14.4 Å for OH:Mg=1.75 (Fig. 4B). Heat-desiccated NAu-2 nontronite samples measured under N2(g) on the CheMin IV demonstrate that the chloritized samples with higher degrees of partial chloritization (i.e. higher OH:Mg ratios) have the largest basal spacings, and those with lower ratios have the smallest basal spacings (Fig. 4C). The OH:Mg=0.5 sample was the least expanded with a d 001 of 9.7 Å, the OH:Mg=1 sample has a d 001 of 14.1 Å, and the OH:Mg=1.5 and 1.75 samples had d 001 of 14.3 Å. Chloritized samples with OH:Mg=1.0, 1.5, and 1.75 also had a peak at ~3.6 Å, consistent with the (004) reflection. The occurrence of this peak along with the (002) at ~7.3 Å demonstrates further a greater degree of chloritization.
Figure 4. XRD patterns from 5 to 30°2θ (20.5 to 3.5 Å) of Mg-saturated and chloritized NAu-2 nontronite. LP-corrected d 001 of Mg-saturated samples and d 001, d 002, d 02l, and d 004 of OH:Mg=1.75 samples are labeled for reference. (A) Patterns measured on the Panalytical at 90% RH. (B) Patterns measured on the Panalytical at ~1% RH (i.e. under dry N2(g)). (C) Patterns of heat-treated samples measured on the CheMin IV under dry N2(g). *The Mg-saturated sample was measured on a Terra instrument under ambient laboratory conditions after heat desiccation.
SWy-1 montmorillonite
All SWy-1 samples have a (06l) peak at ~1.50 Å and a (02l) peak at ~4.48 Å (Fig. 5; see also Table S1 and Fig. S3 in the Supplementary material), which are diagnostic of dioctahedral smectite (e.g. Brindley, Reference Brindley1952; Moore and Reynolds, Reference Moore and Reynolds1997; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014). When measured at 90% RH, the OH:Mg=0.5 sample has the largest basal spacing with a d 001 of 15.8 Å, followed by the Mg-saturated sample with a d 001 of 15.3 Å (Fig. 5A). The chloritized samples with greater OH:Mg ratios have smaller basal spacings than the Mg-saturated sample, where the OH:Mg=1.0, 1.5, and 1.75 samples have d 001 of 13.5, 14.9, and 14.7 Å, respectively. When measured under dry N2(g) on the Panalytical, Mg-saturated montmorillonite has a d 001 of 13.8 Å, and the chloritized samples show an increase in d 001 with an increase in chloritization, from 13.0 Å for the OH:Mg=0.5 sample up to 14.3 Å for the OH:Mg=1.75 sample (Fig. 5B). SWy-1 samples heat-desiccated and measured under N2(g) on CheMin IV have a large range of (001) peak positions. The Mg-saturated sample has a d 001 at 13.9 Å, and, as for the samples measured under dry N2(g) on the Panalytical, the chloritized samples measured on the CheMin IV show an increase in d 001 with increase in chloritization, from 9.8 Å for the OH:Mg=0.5 sample up to 14.4 Å for the OH:Mg=1.75 sample (Fig. 5C). A peak at ~4.9 Å is generally more pronounced in chloritized samples with OH:Mg=1.0, 1.5, and 1.75, but this peak is also present in Mg-saturated patterns. This is the (003) peak, which is higher in amplitude because of the lower octahedral Fe content of SWy-1 compared with the other smectites used in this study. A weak peak at ~3.7 Å is apparent in the OH:Mg=1.75 sample, in particular, which is consistent with the (004) reflection.
Figure 5. XRD patterns from 5 to 30°2θ (20.5 to 3.5 Å) of Mg-saturated and chloritized SWy-1 montmorillonite. LP-corrected d 001 of Mg-saturated samples and d 001, d 002, d 02l, and d 004 of OH:Mg=1.75 samples are labeled for reference. (A) Patterns measured on the Panalytical at 90% RH. (B) Patterns measured on the Panalytical at ~1% RH (i.e. under dry N2(g)). (C) Patterns of heat-treated samples measured on the CheMin IV under dry N2(g).
Thermal and evolved-gas analysis
GP saponite
All Mg-saturated and chloritized GP saponite samples evolved H2O (m/z=18) with peaks between 106 and 183°C due to adsorbed and interlayer H2O, between 439 and 453°C from Fe-OH dehydroxylation in the octahedral sheet, and between 672 and 762°C from Mg-OH dehydroxylation in the octahedral sheet (Fig. 6; see also Table S2 in the Supplementary material; e.g. McAdam et al., Reference McAdam, Sutter, Archer, Franz, Wong, Lewis, Eigenbrode, Stern, Knudson, Clark, Andrejkovicova Slavka Ming, Morris, Achilles, Rampe, Bristow, Navarro-Gonzalez, Mahaffy, Thompson, Gellert, Williams, House and Johnson2020). Chloritized samples with OH:Mg ratios of 1.5 and 1.75 evolved additional sharp water releases with peaks at 328°C and 339°C, respectively, from the breakdown of interlayer Mg(OH)2. Low- and mid-temperature water releases co-occurred with DSC endotherms and TG weight losses (Fig. 6A). All Mg-saturated and chloritized GP samples had DSC endotherms immediately followed by exotherms between 755 and 850°C, caused by phyllosilicate dehydroxylation and recrystallization (e.g. Che et al., Reference Che, Glotch, Bish, Michalski and Xu2011; McAdam et al., Reference McAdam, Sutter, Archer, Franz, Wong, Lewis, Eigenbrode, Stern, Knudson, Clark, Andrejkovicova Slavka Ming, Morris, Achilles, Rampe, Bristow, Navarro-Gonzalez, Mahaffy, Thompson, Gellert, Williams, House and Johnson2020). Water release peaks from 672–710°C decrease in intensity with chloritization, possibly because they are partially obscured by mid-temperature water-release peaks. TG weight loss during pyrolysis generally increased with increasing chloritization. Mg-saturated GP lost 8.6 wt.%, and the chloritized samples lost a total of ~16–18 wt.% (Fig. 6). The greatest difference in TG weight loss between samples occurred over the range 245–725°C.
Figure 6. Mass-corrected thermal and evolved gas data for GP saponite samples. Vertical dashed lines are at 105°C, 335°C, 450°C, and 725°C and represent approximate temperatures of major water peaks. Top: EGA traces for m/z=18 (i.e. H2O). ic = ion current (amps). Middle: baseline-corrected DSC data. Reduced signal of the Mg-saturated sample compared with chloritized samples may be a result of a clog in the mass spectrometer orifice when that sample was analyzed. Bottom: TG data, showing mass loss with increase in temperature.
NAu-2 nontronite
All Mg-saturated and Mg(OH)2 chloritized NAu-2 samples evolved low-temperature adsorbed and interlayer H2O with peaks between ~85°C and 120°C as well as mid-temperature water releases with peaks between ~405°C and 450°C (Fig. 7; see also Table S2 in the Supplementary material). Mg-saturated and chloritized samples evolved mid-temperature water releases with similar peak temperatures, probably caused by the co-occurrence of dehydroxylation of the nontronite octahedral sheet (~450°C; e.g. Heller-Kallai and Rozenson, Reference Heller‐Kallai and Rozenson1980; Che et al., Reference Che, Glotch, Bish, Michalski and Xu2011; McAdam et al., 2020) and the breakdown of interlayer Mg(OH)2. Low- and mid-temperature water releases co-occurred with DSC endotherms and TG weight losses (Fig. 7). Generally, chloritized samples with greater chloritization (i.e. higher OH:Mg ratios) lost more weight during pyrolysis, and the differences in weight loss occurred over the range ~325–560°C. Total weight loss values from the chloritized NAu-2 nontronite ranged from ~11 to 17 wt.%, whereas Mg-saturated NAu-2 lost ~8 wt.%.
Figure 7. Mass-corrected thermal and evolved gas data for NAu-2 nontronite samples. Vertical dashed lines are at 170°C and 415°C and represent approximate temperatures of major water peaks. Top: EGA traces for m/z=18 (i.e. H2O). ic = ion current (amps). Reduced signal of the Mg-saturated sample compared with chloritized samples may be a result of a clog in the mass spectrometer orifice when that sample was analyzed. Middle: baseline-corrected DSC data. Bottom: TG data, showing mass loss with increase in temperature.
SWy-1 montmorillonite
SWy-1 chloritized samples evolved up to four distinct water-release peaks, whereas Mg-saturated SWy-1 only evolved adsorbed water with a peak at ~110°C and water from the dehydroxylation of the octahedral sheet with a peak at 715°C (Fig. 8; see also Table S2 in the Supplementary material). All chloritized SWy-1 samples evolved mid-temperature water with peaks between 440°C and ~450°C, and the sample with OH:Mg=0.5 had an additional peak at ~580°C. Mid-temperature water releases are consistent with the breakdown of Mg(OH)2 in the interlayer site. Mg-saturated SWy-1 and the chloritized SWy-1 sample with OH:Mg=0.5 evolved water with peaks between ~705°C and 715°C from dehydroxylation of the montmorillonite octahedral sheet (e.g. Heller-Kallai and Rozenson, Reference Heller‐Kallai and Rozenson1980; Muller et al., Reference Muller, Drits, Plancon and Robert2000; Che et al., Reference Che, Glotch, Bish, Michalski and Xu2011; McAdam et al., Reference McAdam, Sutter, Archer, Franz, Wong, Lewis, Eigenbrode, Stern, Knudson, Clark, Andrejkovicova Slavka Ming, Morris, Achilles, Rampe, Bristow, Navarro-Gonzalez, Mahaffy, Thompson, Gellert, Williams, House and Johnson2020). The OH:Mg=0.5 sample also evolved water with a peak at ~855°C. The low- and mid-temperature water releases co-occurred with DSC endotherms. All Mg-saturated and chloritized SWy-1 samples had high-temperature (>700°C) endotherms immediately followed by weak exotherms, which were caused by dehydroxylation of the octahedral sheet and subsequent recrystallization (e.g. Che et al., Reference Che, Glotch, Bish, Michalski and Xu2011). Mg-saturated SWy-1 lost a total mass of ~9.4 wt.%, and this value generally increased in samples with higher OH:Mg ratios. The differences in TG mass loss between SWy-1 samples occurred primarily over the temperature range of ~370–670°C, probably related to differences in the extent of chloritization.
Figure 8. Thermal and evolved gas data for SWy-1 montmorillonite samples. Data are mass-corrected. Vertical dashed lines are at 170°C, 445°C, 575°C, and 715°C and represent approximate temperatures of major water peaks. Top: EGA traces for m/z=18 (i.e. H2O). ic = ion current (amps). Reduced signal of the Mg-saturated sample compared with chloritized samples may be a result of a clog in the mass spectrometer orifice when that sample was analyzed. Middle: baseline-corrected DSC data. Bottom: TG data, showing mass loss with increase in temperature.
VSWIR spectroscopy
GP saponite
VSWIR spectra of Mg-saturated and chloritized GP saponite measured under ambient conditions and at 25°C under N2(g) displayed bands from combinations and overtones of fundamental H2O stretching and bending modes at 1.39–1.41 μm and 1.91–1.92 μm and from Mg/Fe-OH combination absorptions at ~2.32–2.33 μm, typical of Fe-bearing saponite (Fig. 9; see also Tables S3–S5 in the Supplementary material) (e.g. Clark et al., Reference Clark, King, Klewja, Swayze and Vergo1990; Ehlmann et al., Reference Ehlmann, Bish, Ruff and Mustard2012; Fox et al., Reference Fox, Kupper, Ehlmann, Catalano, Razzell-Hollis, Abbey, Schild, Nickerson, Peters, Katz and White2021). The ~1.4 μm band was very weak and the ~1.9 μm band was not apparent in samples measured under N2(g) and heated to 200°C because the latter is derived from the fundamental bending vibrations of the H2O molecule, and H2O is not present in completely collapsed smectite. Band minima do not shift significantly with sample dehydration. Chloritized samples with OH:Mg=0.5 and 1 had weak bands at ~2.42 μm, and chloritized samples with OH:Mg=1.5 and 1.75 had weak bands at ~2.48 μm, both the result of combinations of OH stretching and bending modes. The band at ~2.48 μm appeared as a shoulder in the samples measured under ambient conditions, but became a distinct band when measured under N2(g) at 25°C and 200°C. All spectra of chloritized saponite also had a very weak band at ~2.1 μm that was not apparent in the Mg-saturated spectra. The ~2.1 μm band depth increased with degree of chloritization (i.e. with increase in OH:Mg ratio) (Fig. 10). Weak bands observed at ~2.1 μm in phyllosilicates have been attributed to Mg-OH, Al-OH, and Fe-OH combinations (e.g. Cloutis et al., 2008). The overall low albedo of the Mg-saturated sample relative to the other three samples is attributed to insufficient sample to completely fill the instrument beam diameter on an optically black substrate (see the Methods section).
Figure 9. VSWIR reflectance spectra of Mg-saturated and chloritized GP saponite samples. Vertical dashed lines are at 1.40 μm (a), 1.90 μm (b), 2.10 μm (c), 2.30 μm (d), 2.42 μm (e), and 2.48 μm (f). (A,D) VSWIR spectra of GP samples measured under ambient conditions. (B,E) VSWIR reflectance spectra of GP samples measured under dry N2(g) and 25°C. (C,F) VSWIR reflectance spectra of GP samples measured under dry N2(g) and 200°C.
Figure 10. Relationship between the ~2.1 μm band depth and degree of chloritization. (A) ~2.1 μm band depth versus OH:Mg molar ratio for GP saponite measured in air. (B) ~2.1 μm band depth versus OH:Mg molar ratio for NAu-2 nontronite measured in air. (C) ~2.1 μm band depth versus OH:Mg molar ratio for SWy-1 montmorillonite measured in air.
Very small amounts of Fe and changes in coordination and charge have significant effects on the visible/near-infrared portion of the spectra via electronic absorptions (e.g. Rossman and Ehlmann, Reference Rossman, Ehlmann, Bishop, Bell and Moersch2019). The most prominent characteristic in the visible is an oxygen-metal charge transfer absorption, related to Fe3+ and centered in the ultraviolet, therefore creating a reddish edge. This feature shifts to shorter wavelengths with lesser charge, and reduction of some iron during N2(g) purge with heating might drive the diminishment of the red edge. This change was most dramatic in the Mg-saturated sample where the interlayer was most collapsed. The Mg-saturated sample was noticeably darker than those that had undergone chloritization. Speculatively, intervalence charge transfer due to Fe2+–Fe3+, which tends to subdue absorptions over 0.5–1.0 μm (as in magnetite), may be more prevalent in the Mg-saturated sample before further treatment renders samples more ferric. Relatively subtle super-imposed absorptions at ~0.5, 0.65, 0.8, and 0.9 μm were seen in the visible/near-infrared and are consistent with octahedrally coordinated Fe. They do not appear to follow a pattern with treatment and may be due to subtle differences in sample phyllosilicates and nanophase iron oxides, as also seen in subsamples of Griffith saponite (Treiman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014).
NAu-2 nontronite
The VSWIR spectra of Mg-saturated and chloritized NAu-2 nontronite samples measured under ambient conditions and under N2(g) demonstrate H2O stretching and bending modes at ~1.4 and at ~1.9 μm and from Mg-OH combination absorptions at ~2.3 μm (Fig. 11; see also Tables S3–S5 in the Supplementary material) are consistent with nontronite spectra (e.g. Clark et al., Reference Clark, King, Klewja, Swayze and Vergo1990; Bishop et al., Reference Bishop, Murad and Dyar2002; Ehlmann et al., Reference Ehlmann, Bish, Ruff and Mustard2012; Fox et al., Reference Fox, Kupper, Ehlmann, Catalano, Razzell-Hollis, Abbey, Schild, Nickerson, Peters, Katz and White2021). The band at ~1.4 μm is a doublet with minima at 1.39 and 1.43 μm. The ~1.4 μm and ~1.9 μm bands typically weakened with dehydration of samples (i.e. from measurements in air to N2(g) at 25°C to N2(g) at 200°C). Band minima did not shift significantly with sample dehydration. The Mg-saturated and chloritized samples had Mg/Fe-OH combination bands at ~2.3, ~2.4, and ~2.5 μm in all spectra, from combinations of OH stretching and bending modes. All spectra contained very weak bands at ~2.1 μm, which have been observed in phyllosilicates previously and attributed to Mg-OH, Al-OH, and Fe-OH combinations (e.g. Cloutis et al., 2008). The ~2.1 μm band depth increased with degree of chloritization (i.e. with increase in OH:Mg ratio) (Fig. 10). The overall low albedo of the Mg-saturated samples relative to the other three samples is attributed to insufficient sample to completely fill the instrument beam diameter on an optically black substrate (see the Methods section).
Figure 11. VSWIR reflectance spectra of Mg-saturated and chloritized NAu-2 nontronite. Vertical dashed lines are at 1.39 μm (a), 1.43 μm (b), 1.91 μm (c), 2.10 μm (d), 2.29 μm (e), 2.40 μm (f), and 2.50 μm (g). (A,D) VSWIR spectra of NAu-2 samples measured under ambient conditions. (B,E) VSWIR reflectance spectra of NAu-2 samples measured under dry N2(g) and 25°C. (C,F) VSWIR reflectance spectra of NAu-2 samples measured under dry N2(g) and 200°C.
Spectra in the visible/near-infrared displayed a prominent red edge from Fe3+-oxygen charge transfer in the UV and super-imposed absorptions from electronic transitions of octahedrally coordinated iron at ~0.5, 0.65, and 0.9 μm. As with the saponite, the samples exhibited a diminished red edge following heating and N2(g) purge.
SWy-1 montmorillonite
The VSWIR spectra of Mg-saturated and chloritized NAu-2 nontronite samples measured under ambient conditions and under N2(g) demonstrate OH and H2O stretching and bending modes at ~1.4 μm and ~1.9 μm and Mg/Fe(III)/Al-OH combination absorptions at ~2.1 μm to ~2.4 μm (Fig. 12; see also Tables S3–S5 in the Supplementary material), consistent with Mg-bearing montmorillonite (e.g. Clark et al., Reference Clark, King, Klewja, Swayze and Vergo1990; Bishop et al., Reference Bishop, Murad and Dyar2002; Ehlmann et al., Reference Ehlmann, Bish, Ruff and Mustard2012; Fox et al., Reference Fox, Kupper, Ehlmann, Catalano, Razzell-Hollis, Abbey, Schild, Nickerson, Peters, Katz and White2021). The band minima at ~1.4 μm and ~1.9 μm weakened with dehydration (i.e. from measurements in laboratory air to N2(g) at 25°C to N2(g) at 200°C) in all chloritized and Mg-saturated samples. Samples with OH:Mg=1, 1.5, 1.75 had an additional band at ~2.5 μm when measured under ambient conditions and N2(g) at 25°C. Under N2(g) 200°C, only the chloritized samples with OH:Mg=1.5 and 1.75 had an additional band at ~2.5 μm, which is a combination of OH stretching and bending vibrations.
Figure 12. VSWIR reflectance spectra of Mg-saturated and chloritized SWy-1 montmorillonite. Vertical dashed lines are at 1.39 μm (a), 1.41 μm (b), 1.91 μm (c), 2.09 μm (d), 2.21 μm (e), 2.35 μm (f), 2.44 μm (g), and 2.48 μm (h). (A,D) VSWIR spectra of SWy-1 samples measured under ambient conditions. (B,E) VSWIR reflectance spectra of SWy-1 samples measured under dry N2(g) and 25°C. (C,F) VSWIR reflectance spectra of SWy-1 samples measured under dry N2(g) and 200°C.
Spectra in the visible/near-infrared displayed a weak absorption edge possibly from Fe3+–oxygen charge transfer absorption in the UV that is weaker than the other smectites because of a lower Fe concentration (e.g. Mermut and Faz Cano, Reference Mermut and Faz Cano2001). Weak electronic transitions at ~0.45, ~0.65, and ~0.9 μm are consistent with the presence of minor Fe3+.
Discussion
Effects of partial chloritization on XRD, EGA, and VSWIR data
XRD patterns revealed the extent of dehydration and chloritization with changing OH:Mg ratio and the effects of layer charge on chloritization. The heated samples measured under dry N2(g) on the CheMin IV were more dehydrated than the samples measured on the Panalytical under dry N2(g) based on the smaller d 001 in CheMin IV patterns. At near 0% RH, Mg-saturated smectite is expected to have a d 001 near 12 Å (e.g. Ferrage et al., Reference Ferrage, Lanson, Sakharov and Drits2005), but the Mg-saturated smectite samples measured under dry N2(g) on the Panalytical had d 001 near 13.5 Å. This discrepancy in peak position suggests that the samples did not fully equilibrate in the dry N2(g) on the Panalytical. Mg-saturated samples measured under N2(g) on the CheMin IV dehydrated more consistently to a d 001 of <13.5 Å. Thus, patterns for samples that were heated and measured under N2(g) on the CheMin IV best approximate XRD data collected from the very dry surface of Mars.
The XRD patterns demonstrated that treatments successfully produced chloritized smectite samples with increasing OH:Mg ratio. The chloritized smectites measured under dry N2(g) (i.e. with most interlayer water driven off) on the Panalytical and CheMin IV showed an increase in d 001 from OH:Mg=0.5 to OH:Mg=1.75, indicating the extent of chloritization increased with OH:Mg from more extensive occurrence of brucite-like sheets in the interlayer site. The most dehydrated chloritized smectite samples (i.e. those measured on the CheMin IV under N2(g)) showed substantial variability within the same smectite type with different OH:Mg ratios and between smectite types. The ~10 Å d 001 for Mg-saturated GP saponite indicates that this sample collapsed fully. The 12.9 Å and 13.9 Å d 001 for Mg-saturated NAu-2 nontronite and SWy-1 montmorillonite, respectively, suggest that these samples were not fully dehydrated before analysis. The increase in d 001 to 12.2 Å for GP saponite partially chloritized at OH:Mg=0.5 suggests that this sample was partially chloritized, whereas the d 001 for NAu-2 nontronite and SWy-1 montmorillonite at OH:Mg=0.5 at ~10 Å demonstrates that these samples were not chloritized. This difference in extent of partial chloritization at the lowest OH:Mg may be related to differences in layer charge between the three smectites. GP saponite is a high-charge smectite (with a total layer charge of –0.74), where most of the layer charge is on the tetrahedral sheet (e.g. Vicente et al., Reference Vicente, Bañares-Muñoz, Suárez, Pozas, López-González, Santamaría and Jiménez-López1996). NAu-2 nontronite has a similar layer charge (–0.7 per unit cell; Gates et al., Reference Gates, Slade, Manceau and Lanson2002), where some is tetrahedral (–0.45), and some is octahedral (–0.27). SWy-1 montmorillonite has the lowest layer charge of the three (–0.55), with almost all of the charge associated with the octahedral sheet (e.g. Mermut and Lagaly, Reference Mermut and Lagaly2001). Previous partial chloritization studies on saponite that precipitated gibbsite-like sheets (i.e. Al-OH) in the interlayer site showed that the extent of chloritization is greater for high-charge smectite where the layer charge originates in the tetrahedral sheet (e.g. Kloprogge et al., Reference Kloprogge, Booy, Jansen and Geus1994; Vicente et al., Reference Vicente, Suarez, Bañares-Muñoz and Martin-Pozas1997). The high tetrahedral layer charge on the GP saponite may have promoted more complete chloritization at the lowest OH:Mg, resulting in a higher d 001. All chloritized smectite samples reached their maximum d 001 at OH:Mg=1.75 (13.8 Å, 14.3 Å, and 14.4 Å for GP, Nau-2, and Swy-1, respectively).
Based on established methods (e.g. Moore and Reynolds, Reference Moore and Reynolds1997), NEWMOD software was used to model the basal reflections of the XRD patterns measured on the Panalytical under dry N2(g) and to estimate the extent of chloritization for each sample. The basal reflections were modeled using two phases: (1) discrete collapsed smectite, and (2) mixed-layer chlorite-smectite (C-S). The C-S model provided a valid representation of the experimental products of partial chloritization because the extent to which chlorite interlayer hydroxide sheets are filled can vary. Moreover, the ratio of chlorite-to-smectite is optimized during the modeling procedure, mimicking layer-to-layer differences in susceptibility to chloritization. The hydration state of the smectite layers in C-S was allowed to vary. Model instrument parameters were set to match the experimental set-up (X-ray wavelength = 1.7902 Å, goniometer radius = 24 mm, Søller slit width = 2.3, divergence slit width = 0.5). The degree of chloritization of samples was estimated as the product of the modeled ratio of C-S to discrete smectite, the percentage C content of C-S, and the occupancy of the chlorite interlayer hydroxide sites. For example, the measured XRD pattern of the GP saponite with OH:Mg=0.5 was best fit by a model containing a mixture of 77% C-S and 23% discrete collapsed smectite, with C-S containing 80% chlorite layers, and a chlorite interlayer hydroxide site occupancy of 69%. Thus, resulting in an estimated degree of chloritization of 77%×80%×69%=43% (Table 2; see also Figs S4–S14 in the Supplementary material). NEWMOD-derived estimates of the degree of chloritization of smectite standards all increase with increasing OH:Mg in the chloritizing solutions.
Table 2. Estimated degree of chloritization of laboratory-chloritized smectites and the CheMin Cumberland sample.
Estimates were derived from NEWMOD models of measured XRD patterns. *The NAu-2 OH:Mg 0.5 XRD pattern was too complex to model using the approach outlined in the main text because of the presence of smectite in various hydration states.
NEWMOD was also used to model the basal phyllosilicate reflections of the Cumberland CheMin XRD pattern, using a similar approach. NEWMOD’s background correction function was applied (Fig. 13) to the Cumberland pattern to account for amorphous material contributions – the ‘low-angle rise’ commonly observed in amorphous-bearing samples (e.g. Pandey et al., Reference Pandey, Rampe, Ming, Deng, Bedford and Schwab2023) – that may influence the apparent position of basal reflections. Modeling suggests that the phyllosilicate in this drill target is ~93% mixed-layer chlorite-ferrian saponite and ~7% collapsed ferrian saponite (Fig. 13). The modeled percentage C content of C-S is ~70%, with ~92% of the hydroxide interlayer sites of the chlorite filled by Mg(OH)2. The calculations from these model results suggest that the smectite in Cumberland is ~60% chloritized, similar to smectites exposed to fluids with OH:Mg=1.0 in the laboratory syntheses presented here. This finding indicates that there was insufficient Mg2+ or OH– in solution to fully chloritize the smectite in Cumberland, suggesting further that the interaction between saponite in Yellowknife Bay and alkaline Mg2+-bearing fluids was relatively brief.
Figure 13. Model of degree of chloritization in Cumberland. (A) NEWMOD model (yellow trace) of the basal phyllosilicate reflections in the Cumberland CheMin XRD pattern (red pattern). The yellow trace is a linear combination of a mixed-layer chlorite-smectite (red trace) and a discrete collapsed smectite (pink trace). The model uses ~93 wt.% chlorite-smectite and ~7 wt.% discrete collapsed smectite. (B) Graphical user interface of NEWMOD showing relevant modeled parameters. Hydroxide layer refers to the occupancy of the hydroxide interlayer sheets of the chlorite layers in C-S (92% in this model). Decimal fraction refers to the weight per cent chlorite in the C-S (70% in this model).
EGA data demonstrated that partial chloritization with Mg(OH) results in a distinct H2O release peak near 450°C that may be a defining characteristic of this type of partial chloritization. This peak was apparent in all chloritized samples and increased in intensity with higher OH:Mg ratios (i.e. with greater extent of chloritization), demonstrating that more Mg–OH bonds are broken with pyrolysis. GP saponite samples with the highest OH:Mg (i.e. 1.5 and 1.75) had an additional sharp peak at 335°C, which is consistent with the thermal decomposition of brucite (e.g. Gallagher, Reference Gallagher1982). The H2O release from the breakdown of Mg-OH bonds in the interlayer site of chloritized NAu-2 nontronite samples overlapped the H2O release from the breakdown of Fe(III)-OH bonds in the octahedral site. This overlap indicates EGA cannot be used to definitively identify nontronite chloritized with brucite-like sheets.
Chloritization appears to destabilize the structure of the GP saponite and SWy-1 montmorillonite octahedral sheets somewhat. The Mg-saturated GP saponite EGA data showed an H2O release at ~850°C from the dehydroxylation of the Mg–OH bonds in the octahedral sheet (e.g. McAdam et al., 2020). With increasing degree of chloritization, this release shifted to lower temperatures (~700°C) and became weaker, suggesting that the octahedral sheet is less stable than in the parent saponite and that fewer Mg-OH bonds are available to break. EGA data from SWy-1 montmorillonite showed that the peak at ~715°C from the dehydroxylation of the Al-OH bonds in the octahedral sheet either disappears with increasing degree of partial chloritization or shifts to lower temperatures (~575°C) to account for an asymmetric release at 445°C. The disruption of the octahedral sheet was not observed by XRD (i.e. the (06l) peak positions and intensities did not vary), so this disruption is probably not substantial. Alternatively, partial chloritization with Mg-OH may catalyze the dehydroxylation of saponite and montmorillonite octahedral sheets, where the release of H2O at ~425°C from the breakdown of interlayer brucite-like sheets destabilizes the mineral structure and allows the breakdown of octahedral sheets at lower temperatures. This catalysis does not happen in nontronite because the dehydroxylation of the brucite-like sheets and the Fe–OH bonds in the octahedral sheet occur at the same temperature.
Partial chloritization of GP saponite, NAu-2 nontronite, and SWy-1 saponite had an effect on VSWIR reflectance spectra but did not produce strong diagnostic bands that could be identified easily on the Martian surface from orbit or in situ. The primary vibrational bands used to identify phyllosilicates in the laboratory and on Mars (i.e. M-OH overtones between ~2.2 and ~2.3 μm) were not affected by partial chloritization, further suggesting that the structure of the parent smectite is not substantially changed by partial chloritization. VSWIR laboratory measurements in air and under dry N2(g) at 25°C typically exhibited decreased ~1.9 μm H2O band intensity with an increase in partial chloritization from the replacement of interlayer H2O with interlayer brucite-like sheets. The GP saponite samples chloritized with OH:Mg=1.5 and 1.75 had a weak band at 2.48 μm, which was also seen in spectra of brucite (e.g. Beck et al., Reference Beck, Schmidt, Cloutis and Vernazza2015) and, therefore, is not a unique characteristic of Mg-OH partial chloritization (Fig. 14).
Figure 14. Comparisons of VSWIR reflectance spectra of partially chloritized smectite, Mg-rich chlorite, and brucite. VSWIR reflectance spectra of OH:Mg=1.75 partially chloritized GP saponite (red), NAu-2 nontronite (blue), and SWy-1 montmorillonite (green) measured in air and brucite (black dashed) and Mg-rich chlorite (gray dashed). Brucite and Mg-rich chlorite spectra are from the USGS Spectral Library (Brucite_HS247.1B_ASDFRc_AREF and Chlorite_SMR-13.d_30-45um_BECKa_AREF, respectively) (Kokaly et al., Reference Kokaly, Clark, Swayze, Livo, Hoefen, Pearson, Wise, Benzel, Lowers, Driscoll and Klein2017). Vertical dashed lines are at 1.4 μm (a), 1.91 μm (b), 2.09 μm (c), 2.21 μm (d), 2.3 μm (e), 2.4 μm (f), and 2.48 μm (g) for reference.
The weak band at ~2.1 μm may be a marker of degree of partial chloritization; the intensity of the 2.1 μm band increased with increasing degree of chloritization (Fig. 10). This band was previously recognized in laboratory spectra of a variety of smectite samples and was assigned to Mg-OH, Al-OH, and Fe-OH combinations (e.g. Cloutis et al., Reference Cloutis, Craig, Kruzelecky, Jamroz, Scott, Hawthorne and Mertzman2008). A weak band at ~2.1 μm is also present in brucite (e.g. Beck et al., Reference Beck, Schmidt, Cloutis and Vernazza2015), and thus it is not uniquely attributable to chloritization of smectite (Fig. 14). Additional complications with using this peak to identify chloritized smectite on Mars is that this band is probably too weak to be identified in CRISM data above the noise threshold and is in the same location as some sulfate bands (e.g. Cloutis et al., Reference Cloutis, Craig, Kruzelecky, Jamroz, Scott, Hawthorne and Mertzman2008). The ~2.1 μm band may be useful when used in concert with other datasets on landed missions. For example, VSWIR reflectance spectra from SuperCam on the Mars2020 Perseverance rover (Wiens et al., Reference Wiens, Maurice, Robinson, Nelson, Cais and Bernardi2021) with a signature of smectite with a weak ~2.1 μm band, combined with spatial geochemical data from the Planetary Instrument for X-ray Lithochemistry (PIXL) (Allwood et al., Reference Allwood, Wade, Foote, Elam, Hurowitz and Battel2020) that showed an enrichment in Mg, would be consistent with Mg-OH chloritized smectite, but detailed structural analyses on the clay minerals in returned samples would be necessary to confirm this hypothesis.
Evidence for partial chloritization in Gale crater and implications for aqueous environments
CheMin XRD and SAM EGA data are consistent with the presence of partially chloritized saponite at the base of the stratigraphic section in Yellowknife Bay (Table 3). The CheMin XRD pattern of the Cumberland drill powder shows a (001) peak at ~13.5 Å and a shoulder at 10 Å, and the XRD pattern of the John Klein powder shows a (001) peak at 10 Å with a shoulder at higher d-spacings (Fig. 1; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014). The (02l) peaks of these drill powders and the H2O release at 725°C from SAM data demonstrate that the clay mineral present is Fe3+-bearing saponite (Ming et al., Reference Ming, Archer, Glavin, Eigenbrode, Franz and Sutter2014; Treiman et al., Reference Treiman, Morris, Agresti, Graff, Achilles, Rampe, Bristow, Ming, Blake, Vaniman, Bish, Chipera, Morrison and Downs2014; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014; Bristow et al., Reference Bristow, Bish, Vaniman, Morris, Blake, Grotzinger, Rampe, Crisp, Achilles, Ming, Ehlmann, King, Bridges, Eigenbrode, Sumner, Chipera, Morookian, Treiman, Morrison, Downs, Farmer, Des Marais, Sarrazin, Floyd, Mischna and McAdam2015). Measurements of chloritized smectite under dry N2(g) showed that partial chloritization with Mg-OH can produce expanded (001) peaks under Martian conditions, like those seen in Cumberland and John Klein patterns. It was previously proposed that an interlayer cation with a high hydration energy (e.g. Mg2+) could also cause an expanded smectite structure in Cumberland and John Klein (Ming et al., Reference Ming, Archer, Glavin, Eigenbrode, Franz and Sutter2014; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014; Bristow et al., Reference Bristow, Bish, Vaniman, Morris, Blake, Grotzinger, Rampe, Crisp, Achilles, Ming, Ehlmann, King, Bridges, Eigenbrode, Sumner, Chipera, Morookian, Treiman, Morrison, Downs, Farmer, Des Marais, Sarrazin, Floyd, Mischna and McAdam2015; Fukushi et al., Reference Fukushi, Sekine, Sakuma, Morida and Wordsworth2019). However, the observation here of complete collapse of heat-desiccated Mg-saturated GP saponite measured under N2(g) on the CheMin IV indicates that partial chloritization of saponite is necessary to explain the expanded structures in Cumberland and John Klein. Cumberland has a weak diffraction peak at ~7.4 Å, which was previously attributed to the (101) peak in akaganeite (Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014). This peak may be the (002) peak of partially chloritized smectite in this mudstone.
Table 3. Comparisons of XRD and EGA observations in Cumberland (CB), John Klein (JK), partially chloritized GP saponite, and Mg-saturated GP saponite.
* Laboratory XRD observations are from measurements in the CheMin IV under dry N2(g) after heating.
SAM H2O release data are also consistent with the presence of Mg-OH partially chloritized saponite in Cumberland and John Klein. Fitting peaks in the SAM H2O release data revealed a shoulder near 450°C, which is in the temperature range of interlayer Mg-OH dehydroxylation (Fig. 15). Although nontronite also dehydroxylates at this temperature (Fig. 7), the lack of a (02l) peak diagnostic of a dioctahedral smectite in CheMin data of John Klein and Cumberland indicates that nontronite cannot explain the EGA 450°C peaks from both drill samples. Models of SAM H2O EGA traces showed that ~25% of the total water in John Klein is accounted for by this peak, whereas ~11% of the total water in Cumberland is accounted for by this peak (Table 4). This discrepancy is the result of greater relative amounts of low-temperature water in Cumberland compared with John Klein, which could be associated with interlayer H2O in the smectite structure or H2O adsorbed to X-ray amorphous materials (Ming et al., Reference Ming, Archer, Glavin, Eigenbrode, Franz and Sutter2014).
Figure 15. Fitting SAM H2O EGA traces from John Klein (A) and Cumberland (B). Yellow traces show the original SAM data and purple curves are the cumulative fits from the four individual H2O peaks identified in the models. SAM data from John Klein and Cumberland show shoulders near 450°C that are modeled with peaks shown as blue curves.
Table 4. H2O associated with the ~450°C peak in SAM data from John Klein (JK) and Cumberland (CB)
The presence of saponite partially chloritized with Mg(OH)2 in Yellowknife Bay is consistent with limited interaction with alkaline Mg-bearing fluids. Raised ridges enriched in Mg observed in the bedrock that hosts the Cumberland and John Klein drill holes suggest these Mg-rich alkaline fluids were present during early diagenesis before the sediments lithified (Léveillé et al., Reference Léveillé, Bridges, Wiens, Mangold, Cousin and Lanza2014). Chloritized smectite is also consistent with thermochemical models that show alteration of Fe-rich basalt in CO2-poor and oxidizing solutions with a pH of ~7.5–12 and can explain the secondary mineral assemblage in Yellowknife Bay (Bridges et al., Reference Bridges, Schwenzer, Leveille, Westall, Wiens, Mangold, Bristow, Edwards and Berger2015). Such alkaline Mg2+-bearing fluids would be expected for fluids buffered by the basaltic sediments observed in Yellowknife Bay (e.g. Tosca and McLennan, Reference Tosca and McLennan2006). Chloritization of saponite by Mg2+-rich alkaline pore waters has been observed in rocks and sediments on Earth. Evaporative, alkaline environments can produce chloritic material with incomplete layer sheets (e.g. Weaver, Reference Weaver1989; Bristow and Milliken, Reference Bristow and Milliken2011). Jurassic-aged lake sediments from the East Berlin formation of the Connecticut Valley contain chloritized smectite but lack salts typical of an alkaline lake evaporation sequence, suggesting chloritization happened during early diagenesis, rather than as a result of evaporation of lake waters at the surface (April, Reference April1981). A lack of abundant salts or morphological features indicative of desiccation at Yellowknife Bay (e.g. McLennan et al., Reference McLennan, Anderson, Bell, Bridges, Calef and Campbell2014; Vaniman et al., Reference Vaniman, Bish, Ming, Bristow, Morris, Blake, Chipera, Treiman, Rampe, Rice, Achilles, Grotzinger, McLennan, Williams, Bell, Newsom, Downs, Maurice, Sarrazin, Yen, Morookian, Farmer, Stack, Milliken, Ehlmann, Sumner, Berger, Crisp, Hurowitz, Anderson, Des, Marais, Stolper, Edgett, Gupta and Spanovich2014) suggests similarly that partial chloritization occurred during early diagenesis rather than during evaporation of the lake waters.
Smectite is observed throughout most of the 600+ m stratigraphic section studied by Curiosity so far, but Yellowknife Bay is the only location where an expanded structure is observed (e.g. Rampe et al., 2020a; Tu et al., Reference Tu, Rampe, Bristow, Thorpe, Clark and Castle2021; Thorpe et al., 2022). The limited extent of partially chloritized smectite in Gale crater suggests that alkaline Mg2+-bearing fluids did not commonly interact with smectite-bearing units. The location of Yellowknife Bay within Gale crater may explain why this mudstone is the only outcrop that shows evidence for partially chloritized smectite across Curiosity’s traverse. Yellowknife Bay is at the base of the stratigraphic section, making it a natural place for diagenetic fluids to collect. As the fluids interacted with the buried lake sediments in a closed system, the pH would be buffered by basaltic primary minerals, releasing basic cations, including Mg2+, and creating alkaline fluids. Open-system alteration at higher elevations investigated by Curiosity (i.e. on the lower slopes of Mount Sharp) would preclude formation of alkaline fluids because of a shorter residence time. Indeed, observation of a change from trioctahedral to dioctahedral smectite going up section indicates a transition from closed- to open-system alteration (Bristow et al., Reference Bristow, Rampe, Achilles, Blake, Chipera, Craig, Crisp, Des Marais, Downs, Gellert, Grotzinger, Gupta, Hazen, Horgan, Hogancamp, Mangold, Mahaffy, McAdam, Ming, Morookian, Morris, Morrison, Treiman, Vaniman, Vasavada and Yen2018).
Putting the results from Yellowknife Bay in context with the mineralogy of the rest of the stratigraphic section demonstrates a diversity of diagenetic conditions in Gale crater. The observation of variable amounts of akaganeite, jarosite, red and gray hematite, magnetite, and carbonate suggests diagenetic fluids varied in pH, salinity, redox conditions, and temperature (e.g. Rampe et al., Reference Rampe, Ming, Blake, Bristow, Chipera and Grotzinger2017; Rampe et al., 2020a; Rampe et al., Reference Rampe, Bristow, Morris, Morrison, Achilles and Ming2020b; Bristow et al., Reference Bristow, Grotzinger, Rampe, Cuadros, Chipera, Downs, Fedo, Frydenvang, McAdam, Achilles, Blake, Castle, Craig, Des Marais, Downs, Hazen, Ming, Morris, Morrison, Thorpe, Treiman, Tu, Vaniman, Yen, Gellert, Mahaffy, Wiens, Bryk, Bennett, Fox, Milliken, Fraeman and Vasavada2021; Thorpe et al., 2022). The recent orbital and in situ detections of hydrated Mg-sulfate salts in units above the smectite-bearing strata (e.g. Sheppard et al., Reference Sheppard, Milliken, Parente and Itoh2020; Chipera et al., Reference Chipera, Vaniman, Rampe, Bristow, Martínez, Tu, Peretyazhko, Yen, Gellert, Berger, Rapin, Morris, Ming, Thompson, Simpson, Achilles, Tutolo, Downs, Fraeman, Fischer, Blake, Treiman, Morrison, Thorpe, Downs, Castle, Craig, Des Marais, Hazen, Vasavada, Hausrath, Sarrazin and Grotzinger2023) and the detection of siderite (FeCO3) in situ (e.g. Morrison et al., Reference Morrison, Blake, Bristow, Castle, Chipera and Craig2024) suggests that saline Mg2+- and HCO3–-bearing fluids may have been more common later in Gale crater’s history. The absence of smectite in these rocks, however, precluded partial chloritization reactions.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/cmn.2024.42.
Data availability statement
Laboratory data presented in this manuscript are available in the supplementary material. Data from CheMin and SAM are available in the Planetary Data System (https://pds-geosciences.wustl.edu/msl/). CheMin data are also available in the Open Data Repository (http://odr.io/CheMin).
Acknowledgements
The authors thank Michael Thorpe for the use of the NEWMOD software, and Brad Sutter for his thoughtful comments on the manuscript.
Author contribution
EBR conceived the project, designed the chloritization experiments, made laboratory measurements, interpreted laboratory and CheMin data, and wrote the manuscript. VMT helped design the chloritization experiments, made laboratory XRD measurements, interpreted laboratory XRD and VSWIR data, and helped write the manuscript. RVM collected, processed, and interpreted laboratory VSWIR data and edited the manuscript.JVC collected and interpreted laboratory EGA data and helped write and edit the manuscript.TFB interpreted laboratory XRD data, performed NEWMOD analyses, and helped write and edit the manuscript. BLE interpreted laboratory VSWIR data and helped write and edit the manuscript. SP performed chloritization experiments and collected XRD data. VC performed chloritization experiments and collected XRD data. BR interpreted laboratory VSWIR data. DWM helped conceive of the project and edited the manuscript. PDA helped conceive of the project and edited the manuscript.
Financial support
This work was funded by the NASA Solar System Workings Program, grant #80NSSC18K1535, and by the ISFM Mission Enabling Science Work Package at the Johnson Space Center.
Competing interest
The authors declare none.
Open access
