Geochemical biosignature formation in experimental Martian fluvio‐lacustrine and simulated evaporitic settings

To assess whether life existed on Mars, it is crucial to identify geochemical biosignatures that are relevant to specific Martian environments. In this paper, thermochemical modeling was used to investigate fluid chemistries and secondary minerals that would have evolved biotically over geological time scales in Martian fluvio‐lacustrine and evaporitic settings, and that could be used as potential inorganic biosignatures for life detection on Mars. Modeling was performed using fluid and rock chemistries relevant to Gale crater aqueous environments. Potential inorganic biosignatures were identified investigating alteration deposits found at the surface of a simulant exposed to short‐term bio‐mediated weathering and comparing experimental and modeling results. In a fluvio‐lacustrine setting (water/rock of 2000–278), models suggest that less complex mineral assemblages form during biotic basalt dissolution and subsequent brine evaporation compared to what would happen in an abiotic system. Mainly nontronite, kaolinite, and quartz form under biotic conditions, whereas celadonite, talc, and goethite would also precipitate abiotically. Quartz, sepiolite, and gypsum would precipitate from the evaporation of fluids evolved biotically, whereas nontronite, talc, zeolite, and gypsum would form in an abiotic evaporitic environment. These results could be used to distinguish products of abiotic and biotic processes, aiding the interpretation of data from Mars exploration missions.

Biosignature formation depends on a variety of processes that microbes can use to extract bio-essential elements from rocks (Banfield et al., 2001;Cady et al., 2003;Hays et al., 2017;McMahon et al., 2018;Summons et al., 2011;Vago et al., 2017;Westall et al., 2015).In basaltic aqueous environments, similar to early Martian fluvio-lacustrine systems, microbes source bio-essential elements from minerals such as olivine, pyroxene, and plagioclase to produce excess protons, low molecular weight organic acids, siderophores and extracellular polysaccharides and enzymes (e.g., Barker & Banfield, 1998;Bennet et al., 2001;Kalinowski et al., 2000;Liermann et al., 2000;Olsson-Francis et al., 2015, 2017;Rogers et al., 1998;Uroz et al., 2009;Vandevivere et al., 1994;Welch & Ullman, 1993;Welch et al., 2002;Wu et al., 2007).These mechanisms enhance basalt weathering promoting mineral dissolution, changing the chemistry of the fluids and producing distinctive secondary mineral assemblages and mineral surface morphologies (Banfield et al., 2001;Cady et al., 2003;Cogliati et al., 2022;Olsson-Francis et al., 2015, 2017); if preserved within the geological record, such mineralogical and geomorphological features can be potentially used as inorganic biosignatures when searching for evidence of past life in aqueous environments, including paleo lakes and rivers on Mars (Hays et al., 2017;Vago et al., 2017;Westall et al., 2015).Other types of biosignature, such as biominerals and mineral surface textures, may be more appropriate for assessing whether life existed on early Mars since they may be more likely preserved in Martian rocks and detectable through in situ measurements by rover missions (Vago et al., 2017).This is a benefit of this type of biosignatures over organic biosignatures, such as organic molecules, which would be more degradable and less likely to be detectable at the surface owing to the detrimental effect of the extreme conditions at the present-day surface of Mars (Dartnell, 2011;Hays et al., 2017;Ten Kate et al., 2005;Vago et al., 2017).
To identify and use specific secondary mineral assemblages and mineral surface textures as biosignatures for life detection, the processes and products of abiotic and biotic basalt weathering have been extensively analyzed, characterized, and compared in the field and in laboratory experiments (Macey et al., 2023;Oelkers & Gislason, 2001;Olsson-Francis et al., 2015, 2017, 2020;Price et al., 2018;Wolff-Boenisch et al., 2006;Wu et al., 2007).However, due to the limitations of these approaches, biosignature formation has only been studied over comparatively short time frames (monthsyears), which restricts the assessment of the mineralogical associations that would precipitate over geological time scales.Processes that occur over these time scales, such as incongruent or complete dissolution of rocks, may have a great impact on fluid chemistries and secondary mineral precipitation (Bridges et al., 2015).This would in turn influence the availability of bio-essential elements and, thus, the formation and the type of biosignatures (e.g., Benzerara et al., 2004Benzerara et al., , 2005;;Welch & Ullman, 1993).

Geochemical biosignature formation
To investigate the impact of microbial activity on a simulated early Martian water-rock systems, and the subsequent formation of inorganic biosignatures over different time scales, thermochemical modeling has been used in conjunction with experimental microbiology (Cogliati et al., 2022;Olsson-Francis et al., 2017).Thermochemical modeling is able to predict secondary mineral assemblages and variations in fluid chemistries by assessing reaction pathways during water-rock interactions and has therefore been widely applied to study alteration processes that can occur over geological time scales in terrestrial and Martian aqueous environments (Bridges & Schwenzer, 2012;Bridges et al., 2015;Catalano, 2013;Filiberto & Schwenzer, 2013;Griffith & Shock, 1997;Melwani Daswani et al., 2016;Ramkissoon et al., 2021;Schwenzer & Kring, 2013;Zolotov & Mironenko, 2007, 2016).Previous studies that combined laboratory experiments and thermochemical modeling to investigate inorganic biosignatures formation in simulated Martian lacustrine-sedimentary systems identified mineralogical, microscopic, and geochemical changes that could be potentially used to distinguish between abiotic and biotic weathering processes that may have occurred on early Mars (Cogliati et al., 2022;Olsson-Francis et al., 2017).These studies suggest also that, over geological time scales, a less complex secondary mineral assemblage forms during biotic dissolution compared to abiotic weathering (Cogliati et al., 2022;Olsson-Francis et al., 2017).However, these experiments often used nonideal matches of fluid and rock chemistries (i.e., the presence of alteration phases in the simulant, differences in the Fe 2+ /Fe 3+ ratios, and elemental composition of the regolith) compared to Gale Crater's environment.These experiments were also potentially influenced by the persistence of nutrients from the inoculum material.These factors could have influenced the weathering processes and reaction pathways, reducing the fidelity of inorganic biosignatures and precluding definitive conclusions.
To better characterize short-term (1 month) microbial weathering processes in the early Gale crater fluvio-lacustrine system and to increase the accuracy of identified inorganic biosignatures that may have formed in this environment, Macey et al. (2023) performed abiotic and microbially mediated simulation experiments using a chemically accurate simulant regolith (OUCM-1, Ramkissoon et al., 2019) and thermochemically modeled Martian groundwater fluids.These were inoculated with chemolithoautotrophic and chemoorganoheterotrophic anaerobic microbes with metabolisms that are models for potential microbial life that may have existed within aqueous environments at Gale crater in the Noachian (Macey et al., 2023).The study concluded that sulfate reducing, acetogenic, and other anaerobic and fermentative bacteria were able to survive across multiple subculturing steps under Martian conditions, producing shifts in chemistry that could be considered putative biosignatures.These biosignatures were suggested as potentially being detectable using techniques employed in remote sensing payloads such as Fourier transform infrared spectroscopy and near-infrared spectroscopy (Bibring et al., 2005).
This paper extends the work done by Macey et al. (2023) by using thermochemical modeling to investigate how fluid chemistries would evolve, and which secondary mineralogical assemblages would form over geological time scale in a chemically accurate Gale crater aqueous environment under simulated abiotic and biotic conditions.To study reaction pathways in a water-rock system similar to the Rocknest basalt at Gale crater, stepwise titration simulations were performed using the composition of the OUCM-1 regolith simulant and the chemical composition of the thermochemically modeled fluids used in Macey et al. (2023).We compared our model results and the experimental results from Macey et al. (2023) to identify chemical and mineralogical signatures that are uniquely produced during microbialmediated basalt weathering and that may be used as inorganic biosignatures for life detection on Mars.Thermochemical modeling was further used to perform evaporation simulations and identify secondary minerals and fluid chemistries that may have evolved in evaporitic environments during seasonal dry episodes or in the late Hesperian, when Mars experienced more arid hydrological regimes.A better understanding of the alteration assemblages that may have formed during the evaporation of fluids resulting from abiotic and biotic weathering of Rocknest basalts could aid the recognition of mineralogy associated with inorganic biosignatures formation.This study therefore has implication for current Mars missions, as well to aid the interpretation of remote sensing data from orbiters, for future life detection missions, and for interpreting the analysis of Martian samples recovered by the future Mars sample return mission.

Characterization of the Simulant after Dissolution Experiments
Abiotic and biotic dissolution experiments in Macey et al. (2023) were performed to investigate inorganic biosignature formation in simulated Martian environments.Biotic dissolution experiments were developed from anoxic sediment from the Pyefleet mudflats, which was collected using 50 cm 9 15 cm cores at low tide from Pyefleet mudflat, Colne Estuary, East Mersea (UK) (51°48 0 N, 0°22 0 E) in May 2018.An enrichment was produced preparing a slurry comprised of 80 g of anoxic sediment and 240 mL of concentrated RN fluids, and using 10 mL of this to inoculate in a growth medium.The medium contained the OUCM-1 simulant and a simulated RN fluids (1:5 ratio of fluid to simulant), as described previously in Macey et al. (2023).The enrichment was incubated at 25°C for 7 months with a headspace of H 2 /CO 2 (80:20) at 1 bar pressure, with subculturing into lower volumes of an increasingly dilute RN medium every 28 days.Abiotic controls were run in parallel to biotic dissolution experiments.The fluid chemistry used during the experiments had a composition closer to proposed Martian water chemistries and was prepared by thermochemically modeling the interaction between pure water and the OUCM-1 simulant using the program CHIM-XPT (Ramkissoon et al., 2021).This produced an HCO 3À -poor brine which was used to simulate and investigate biosignature formation in a CO 2 -poor aqueous environment.The same fluid composition was used as starting chemistry for titration modeling in this study (see Thermochemical Modeling Section).
Field emission gun-secondary emission microscope (FEG-SEM) analysis was conducted on samples that undergone dissolution experiments performed by Macey et al. (2023).This was done to study the morphological characteristics of the OUCM-1 regolith simulant after abiotic and biotic weathering and use such features as potential inorganic biosignatures in addition to secondary minerals predicted by thermochemical models.For the analysis, the simulant was removed from the Wheaton bottles where abiotic and biotic dissolution experiments were carried out, air-dried, and carbon coated (15-20 lm thickness) on aluminum stubs.A ZEISS Supra 55-VP FEG-SEM, which was operated at an accelerating voltage of 15-20 kV and a 8-10 mm working distance, was used and secondary electron images were compared and screened to identify the presence of surface morphologies that may be used as biosignatures.
In stepwise titration simulations, a specific amount of rock was titrated into a constant amount of the starting fluid.The amount of rock reacted with the fluid is represented by the water to rock ratio (W/R) that is the expression of the reaction progress.The lower the W/R the higher is the amount of rock dissolved in the same amount of water.Chemical reactions that may happen under different environmental conditions were simulated by decreasing the (W/R) M , defined here as the water to rock ratio used in the models (e.g., Bridges et al., 2015;Cogliati et al., 2022;Filiberto & Schwenzer, 2013;Olsson-Francis et al., 2017;Reed, 1982).At (W/R) M ≥ 10 6 , it was modeled an environment where a limited amount of rock is dissolved in a large mass of water (e.g., freshwater inflow, fluid percolating in a fracture, a rock surface exposed to regular precipitation).At (W/R) M < 10, it was simulated an environment where a large volume of rock reacts with a limited volume of water (i.e., a stagnant water table with no exchange or porous sediments).Titration modeling was carried out at 25°C and 1 bar pressure, and the input data included the composition of the OUCM-1 regolith simulant (Table 1) and of a thermochemically modeled fluid (Table 2).These conditions and chemistries correspond to those used in weathering experiments performed in Macey et al. (2023) where a simulant, based on the chemistry of Rocknest sand shadow at Gale Crater, was chosen to represent a "global" basaltic Mars regolith composition (Blake et al., 2013;Ramkissoon et al., 2019).The models were run over a range of (W/R) M between 10 6 and two different pH conditions: In the first model, the pH was treated as free parameter and allowed to vary in response to the interaction between the rock and the fluid as it happens under abiotic conditions.In the second model, the pH was fixed at 7.5 to reproduce the experimental conditions during microbial growth, simulating the progression of reaction pathways in the presence of microbial growth.
Fractional evaporation simulations were performed to investigate mineralogical associations and brine chemistries that would evolve when water was isothermally removed from the system during evaporation, and secondary minerals were fractionated by precipitation and/or isolated by sedimentary processes after their formation.The evaporation models were run at 25°C, 1 bar, using a starting fluid chemistry taken at a (W/R) M of 1000 from previous abiotic and biotic models (Table 2).This (W/R) M step presents the water-rock interval at which the biotic model and bio-mediated weathering experiments performed in Macey et al. (2023) are comparable (see Thermochemical Modeling of Microbial Weathering section).During evaporation modeling, formed minerals were removed from the system after each evaporation run, impeding the precipitated phases rereacting with the fluid after their formation.During evaporation simulations, the pH was set as a free parameter since we do not have specific constraints from laboratory experiments.In this regard, we note that the pH does not deviate from being near-neutral (range 7.5-7.9)when simulating the evaporation of a biotically evolved brine.Repeated simulations with a pH fixed at 7.5 generated identical fluid chemistries and mineral associations.In our evaporation models, the ionic strength varies between 8.5E À3 and 6.2E À1 , and between 1.3E À2 and 1.2 in simulations of the abiotically and biotically evolved brines, respectively.These values are below 3, which is the limit that beyond which the computed activity coefficients are not reliable during modeling, as the code uses the Debeye-Hueckel approach for their calculation (Reed et al., 2010).Thus, our model results can be considered reliable and used to interpret the evolution of aqueous systems on Mars.General information on the code, techniques, application, and limitations can be found in Palandri and Reed (2004), Reed and Spycher (2006), and Reed et al. (2010); details relevant for and specific to this study including the database used for modeling, its limitation, and a list of minerals excluded from the models can be found in the supporting information.

Secondary Alteration Minerals
FEG-SEM was used to investigate morphological features on the surface of the simulant.Little physical or chemical weathering was observed when abiotic samples were analyzed (Figure 1A).Analysis of the samples from the biotic experiments identified the presence of microbial cells, with microbial attachment observed.Samples from the biotic experiments also showed extensive evidence of chemical and morphological changes on the mineral surfaces (Figure 1B-I).

Thermochemical Modeling of Basalt Dissolution
To identify shifts in fluid chemistry and mineral formation that would occur over geological time scales under biotic and abiotic conditions as weathering reactions advance, reaction pathway models were run over a range of (W/R) M between 10 6 and 1. Figures 2  and 3 show the mineralogical associations and fluid compositions predicted by the models.Details of secondary minerals and fluid chemistries of modeled water-rock interaction are reported in the supplementary file.

Abiotic Model
The abiotic model (Figure 2b) forms the baseline to which the biotic models are compared.At a (W/R) M of between 10 6 and ~27,000, the main precipitates were quartz and nontronite.In this (W/R) M range, quartz decreased from ~98 wt% to 54 wt%, while nontronite increased up to 26 wt%; talc and kaolinite were also present as minor phases with abundances <10 wt%.Between (W/R) M of 27,000 and 5000, quartz decreased until it was no longer stable, nontronite increased and Note: Species in the fluid are summarized in the table as one species, but during the modeling were partitioned into several dissolved species.
In the abiotic model, the pH increased from 8.3 to 13.8 over the course of the titration range (Figure 3a).The concentration of most of the dissolved elements and component species varied by, at least, two orders of magnitude, except for NH 4 + which was more stable and less variable (Figure 3a).Mn 2+ , and Na + increased in concentration while Cl À , Mg 2+ , HPO 4 2À , and K + generally decreased.Fe 2+ had a more complex pattern: It increased between (W/R) M of 10 6 and 330,000, then decreased up to (W/R) M of 10,000, and increased again at lower (W/R) M (Figure 3b).

Biotic Model
In the biotic model (Figure 2c), quartz formed at a (W/R) M of 10 6 with 100 wt% abundance, then progressively decreased to <5 wt% over decreasing (W/R) M , and at (W/R) M of 30 was no longer stable.This reduction in quartz coincided with nontronite forming and the (W/R) M decreasing.Nontronite reached its maximum of ~65 wt% at (W/R) M of 100 when talc started to form (>5-18 wt%).Kaolinite was stable throughout the entire range, with its abundance increasing up to 23 wt% at (W/R) M of 100, which then remained stable.Trace amounts (<5 wt%) of apatite, pyrite, goethite, and gypsum were also formed at different (W/R) M .
When the pH was set at 7.5, most of the dissolved elements and component species varied in concentration by at least two orders of magnitude except for H 4 SiO 4 , Al 3+ , and NH 4 + , which were relatively stable (Figure 3c).Only HPO 4 2À and HS À decreased over the course of the titration range, while Cl À , SO 4 2À , Ca 2+ , Mg 2+ , Fe 2+ , K + , Na + , and Mn 2+ increased (Figure 3b).

Thermochemical Modeling of Brine Evaporation
The evaporation of fluids was performed with starting chemistries taken from abiotic and biotic models at (W/R) M of 1000 to identify the evolved mineralogical assemblages (Figure 4) and fluid chemistries (Figure 5).

Abiotic Model
The evaporation of a fluid derived from the abiotic model (Figure 4b) precipitated mainly nontronite (93 wt %) when <1% of water evaporated; talc (100-92 wt%) became the main phase in association with lower amount of nontronite (max 15 wt%) when >80% of water was evaporated; talc was the only precipitant when water was between <80% and 40%; further evaporations of water precipitated analcime; gypsum formed only when the system was almost completely dry (water evaporated The pH in the model varied between 11.2 and 12.9 (Figure 5a) as the amount of water present in the system decreased (Figure 5b).Al 3+ , Mg 2+ , and HPO 4 2À decreased by at least of one order of magnitude, while the other elements remained stable.Such variations are associated with the uptake of specific elements by minerals that precipitated during evaporation.

Biotic Model
The evaporation of fluids derived from the biotic model produced nontronite as the major phase only when >99% of water was present in the system (Figure 4c); quartz was the main secondary phase for most of the evaporation range; sepiolite precipitated together with quartz in models where more than 80% of water evaporated, and gypsum was stable only at very dry conditions (water evaporated >99%).Apatite and pyrite formed only in negligible amount (<<1 wt%).
During the evaporation of the evolved biotically fluids, the pH remained near neutral (7.5-7.9)(Figure 5c).H 4 SiO 4 , Al 3+ , HPO 4 2À decreased by more than one order of magnitude during evaporation, while all the other elements remained stable or had a limited decrease (e.g., SO 4 2À ) (Figure 5d).Such variations are associated with the uptake of specific elements by the minerals that precipitated during evaporation.

Regolith Simulant Dissolution
In this study, the surface of the simulant regolith was screened for secondary alteration products and alteration morphologies with FEG-SEM analysis to identify morphological biosignatures produced by the anaerobic microbes analog to those that may have lived in near-neutral water bodies at Gale and Jezero craters on Noachian Mars (Macey et al., 2023).Analysis of the simulant after biotic weathering experiments showed evidence of direct microbial attachment to the mineral surface (Figure 1B), which means that rock dissolution has occurred through direct contact of the microbes with the regolith simulant; however, it is not possible to exclude that elemental release happened also indirectly (Banfield et al., 2001;Gadd, 2010;Uroz et al., 2009;Vandevivere et al., 1994;Welch et al., 2002).
SEM analysis also identified clear mineralogical and morphological changes specific to the biotic experiments that are interpreted as biosignatures, as nothing comparable was observed in the abiotic controls.Dissolution pits, morphological steps observed along mineral cleavage plains together with other secondary morphologies, and deposits identified on the surfaces of the regolith grains (Figure 1C-I) (i.e., isolated and groups of prismatic crystals, amorphous deposits, clusters of protruding spherical cauliflower-like structures, towerlike structures, sheet, and lamina structures) are all evidence of enhanced mineral dissolution and secondary mineral formation following microbial action (Benzerara et al., 2004(Benzerara et al., , 2005;;Brantley & Chen, 1995;Cogliati et al., 2022;Olsson-Francis et al., 2017;Wu et al., 2007).This unique, diverse, and complex combination of alteration assemblages and morphologies demonstrates that specific microbial metabolisms that was viable under chemically accurate Martian conditions can develop pronounced biosignatures that may be easily detectable in situ using the instruments on board of Martian rovers such as MAHLI (Mars Hand Lens Imager instrument) on the Curiosity rover (Allwood et al., 2020;Edgett et al., 2012), as well as PIXL (Planetary Instrument for X-Ray Lithochemistry), SHERLOC (Scanning Habitable Environments with Raman and Luminescence for Organics and Chemicals), and Wide Angle Topographic Sensor for Operations and eNgineering on Perseverance (Allwood et al., 2020).This is an important achievement considering that previous studies that investigated microbial weathering of a Martian simulant and biosignatures formation using individual microbial strains have shown only limited or rare evidence of alteration mineralogy and morphologies (Cogliati et al., 2022;Olsson-Francis et al., 2017).
These results indicate that the use of a chemically accurate regolith simulant and fluid compositions and microbes with specific chemolithotrophic metabolisms relevant to Mars have increased the likelihood of alteration phase formation.Peeled surfaces with bending laminas and tower-like secondary morphologies identified on regolith simulant that underwent biotic weathering experiments may be phyllosilicate-type minerals with a sheet-like structure, such as those predicted by the biotic thermochemical model.However, a direct comparison between the chemistry of observed secondary deposits and of modeled mineral is not possible.After a first screening, chemical analyses of the simulant undergone weathering experiments were impeded due to the shutdown of the laboratories caused by the Covid-19 pandemic.Multiple investigations performed after 2 years did not detect these specific and diagnostic structures or morphologies at the sample surface, making impossible further chemical investigations on target deposits and, thus, the comparison with modeled results.We suggest that exposure of the samples to the ambient atmosphere may have degraded the samples, contributing to the destruction of the morphologies previously observed.This is a crucial point to consider when handling samples from Mars sample return missions and for future investigation of biosignature formation under laboratory-controlled conditions.
Although a comparison between secondary minerals produced during biotic weathering experiments and modeled secondary assemblages is not possible, the result produced in this study are still useful and may be indicative of mineralogical and morphological features of interest when screening for inorganic biosignatures.They also provide information to support the ongoing sampling activities operated by the Perseverance rover, for planning future life detection missions, and for interpreting the results of the analysis of Martian samples recovered by the future Mars sample return mission (Haltigin et al., 2022;Meyer et al., 2022;Tosca et al., 2022).

Thermochemical Modeling of Microbial Weathering
Thermochemical modeling was used to assist in the identification of possible biosignatures and to determine fluid chemistry variations and secondary minerals that form during microbial weathering of a Martian analog over geological time scales.To compare the experimental and model results by focusing on a specific (W/R) M range, between 2000 and 278 (Figures 2 and 3), which corresponds to the actual (W/R) D interval of the dissolution experiments performed under biotic conditions by Macey et al. (2023).The (W/R) D ratio is defined as the amount of rock simulant dissolved during the experiments; it was assessed by focusing on the most soluble elements (e.g., Na + , K + ) in the fluids resulting from the experiments and calculated as outlined in Olsson-Francis et al. (2017).Na + was specifically selected because it was not incorporated into microbial cells at a detectable level or into any of the minerals predicted to precipitate by the models.The (W/R) M range considered here is also equivalent to a fluvio-lacustrine setting where a limited amount of water interacts with a large volume of rock.Similar environments have been identified at Gale and Jezero craters (Grotzinger et al., 2014;Mangold et al., 2021;Vaniman et al., 2014) where microbial life could have existed during early Mars' history (Rampe, Blake, et al., 2020).With specific reference to the focus of this study, distinct mineral associations and fluid chemistries were observed in the abiotic and biotic models when thermochemically modeled Martian-like fluids were used.
Under biologically mediated conditions, thermochemical modeling predicted the formation of mainly nontronite together with lower amounts of kaolinite followed by quartz, whereas the abiotic system precipitated mainly nontronite and celadonite associated with lower amounts of kaolinite, goethite, and talc (Table 4).Traces of pyrite and apatite precipitate in both abiotic and biotic models.Data from titration models confirm that reaction pathways of basalt-fluid interaction at 1 bar and temperature ≤25°C, which is lower than the habitability threshold of <121°C (e.g., Cockell et al., 2016;Conrad, 2014), produces less chemically and mineralogically complex secondary mineral assemblages under biotic conditions (Cogliati et al., 2022;Olsson-Francis et al., 2017).This is due to the different mineral formation pathways in abiotic and biotic systems.Under biotic conditions, the pH is buffered by microbial action and remains fixed preventing the formation of secondary minerals that would form only at pH values above or below lower than that of the bio-mediated fluid.In an abiotic environment, the pH varies solely in response to water-rock interaction and more minerals form at various pH, increasing the mineralogical diversity in the environment.
Compared to previous studies that used less accurate fluid chemistries and regolith simulant compositions with respect to a Gale crater environment (Cogliati et al., 2022;Olsson-Francis et al., 2017), the models in this study do not predict the formation of high amount of hydrated Al-rich clays (kaolinite >50 wt%), carbonates (mainly siderite), or hydroxides (goethite) under biotic conditions.In this study, the mineralogical association was dominated by phyllosilicates of the smectite group (> 60 wt% nontronite-(CaO 0.5 ) 0.3 Fe 2 3+ (Si, Al) 4 O 10 (OH) 2 ÁnH 2 O) and of the kaolinite-group (kaolinite-Al 2 (Si 2 O 5 )(OH) 4 ).Differences in secondary mineral associations and fluid chemistries between this and previous studies that investigated biosignatures formation through thermochemical modeling (Cogliati et al., 2022;Olsson-Francis et al., 2017) can be due to the variations of the input parameters used for modeling water-rock reaction pathways (fluid and rock chemistry, pH, initial oxidation state) and to the different (W/R) D ranges considered in each individual case (~10 6 to ~10 5 , Olsson-Francis et al., 2017;100 to 38, Cogliati et al., 2022;2000 to 278, this study).
Basalt dissolution experiments performed in a chemically accurate Gale crater aqueous environment for 28 days identified characteristic variations in K + and SO 4 2À concentration between the biotic test group and the abiotic test group (Macey et al., 2023).However, such differences in fluid chemistries have been considered ambiguous biosignatures and are not sufficient alone to indicate the presence of a biogenic signal (Macey et al., 2023).More striking differences in the fluid compositions between abiotic and biotic environments emerge when regolith weathering was simulated over geological time scales through thermochemical modeling.Under simulated alkaline lacustrine-sedimentary conditions, between (W/R) M of 2000 and 278, Na + was observed in similar concentrations (10 À3 moles) in the biotic models as in the growth experiments (Table 3) suggesting that it was used by the microbes only in minimal amounts and that its behavior is similar during short-and long-term biotic weathering.Mn 2+ remains lower in the experimental fluid (10 À6 to 10 À5 moles) compared to model results (10 À5 to 10 À4 moles).This could be related to accumulation of this element in the fluid over time since it would not precipitate in any of the secondary minerals predicted to form biotically during long-term regolith simulant alteration.The concentration of Si, Fe 2+ , Mg 2+ , K + , Ca 2+ , and Al 3+ was higher in the experimental fluids than in the modeled fluids (Table 3), and this suggests that biological activity limits the precipitation of these elements during short-term weathering.This could be due to the use of specific bioessential elements by microbes to form inorganic or organic compounds (e.g., proteins and siderophores) for various processes (i.e., produce polysaccharides and extracellular polymeric substances and production of organic compounds) (Barker & Banfield, 1998;Bennet et al., 2001;Gadd, 2010;Kalinowski et al., 2000;Olsson-Francis et al., 2015;Uroz et al., 2009;Vandevivere et al., 1994;Welch & Ullman, 1993;Wu et al., 2007).These microbially mediated processes could have prevented solubility limits being reached, limiting the formation of secondary minerals that could form during long-term alteration.As predicted by thermochemical models, over geological time scales, various types of Mg-, Ca-, Al-, and Fe-rich phyllosilicates (nontronite, kaolinite) would precipitate in different proportions, in addition to quartz and traces of sulfides (pyrite) and phosphate (apatite) (Figure 2).Differences in the results between this and previous studies (Cogliati et al., 2022;Olsson-Francis et al., 2017), in addition to the reasons listed above (e.g., variation of fluid and rock compositions), could be due to the different types and assemblage of microbes used during dissolution experiments (e.g., individual strains and a microbial community).Different microbial metabolisms can enhance and/or inhibit the rock dissolution and formation of secondary precipitates by altering the concentration of specific elements in the fluid and, thus, the formation of potential geochemical biosignatures (Briggs & Summons, 2014;Price et al., 2018;Tan et al., 2018).
Our models indicate that talc and celadonite precipitate in different proportions together with nontronite and kaolinite during abiotic water-rock interaction.While nontronite and kaolinite formation in Martian lacustrine system is thought to have been a likely process (Bishop et al., 2018;Ehlmann et al., 2009Ehlmann et al., , 2011)), talc on Mars is usually associated with metamorphic settings or high-temperature hydrothermal alteration or burial diagenesis (Bristow et al., 2021;Viviano et al., 2013).However, talc and other talc-like Mg-clays (e.g., kerolite) can be also authigenic in sedimentary deposit (Bristow & Milliken, 2011;Tosca et al., 2011) and, as demonstrated in laboratory tests performed at ambient conditions, it can form as an alteration phase in lowtemperature alkaline aqueous systems (Bricker et al., 1973;Tosca et al., 2011).For this reason, the abiotic formation of talc in Martian fluvio-lacustrine system can be considered plausible.
Celadonite is another secondary phase that precipitates in the abiotic models and that needs further explanation.Contrary to other phyllosilicates, such as nontronite, serpentine, and chlorite that have been found widespread on Mars' surface, celadonite has only been identified from orbit at Mawrth Vallis and Holden crater.At these sites, celadonite was detected within an Fe-rich layer in between Al-phyllosilicates and Fe/Mg smectite bearing units and together with amorphous hydrated Al/Si-rich phases and kaolin-group minerals (Bishop et al., 2008(Bishop et al., , 2013;;Poulet et al., 2014).Rock succession at Mawrth Vallis most likely suggests pedogenic alteration or hydrothermal alteration (T of 20-60°C) of basaltic rocks by groundwater circulation or after deposition into a marine environment (Bishop et al., 2008(Bishop et al., , 2013;;McKeown et al., 2009;Poulet et al., 2014).On Earth, celadonite forms in amygdales, voids, or fractures in mm to lm deposits not only under hydrothermal conditions (e.g., Odin et al., 1988;Weisenberger & Selbekk, 2009) but also as results of direct precipitation from lowtemperature (>17°C) Fe-rich solutions at the sedimentwater interface (Polg ari et al., 2013) and as results of in situ alteration of basalts in local pools of water in continental settings (Singh et al., 2022).Similar processes may have occurred in Gale and Jezero crater lacustrine systems and, for this reason, celadonite on Mars may be more widespread, although difficult to detect.Small amounts of secondary minerals, especially if located in fractures and voids, might be difficult if not impossible to detect and identify by orbiters (e.g., Ehlmann et al., 2011;Murchie et al., 2007), and even by rovers (Blake et al., 2012) due to the difficulty to resolve small mineral abundances.
Considering the results of this study, the detection of secondary phyllosilicates, such as celadonite and talc together with nontronite and kaolinite in Martian samples, could play a role in the distinguishing of mineralogical assemblages produced by abiotic and biotic basalt weathering.In situ analysis by rover missions and/or detailed chemical and mineralogical characterization of returned samples would be the only way to identify specific alteration phases that could be ideal targets for searching traces of extinct and/or extant life on Mars and study its potential habitability.Phyllosilicates such as nontronite and kaolinite would be fundamental for this, as they have a high potential to preserve biosignatures due to their long crustal residence time, low porosity and permeability, and a sheet-like structure that enables adsorption, protection, and preservation of organic matter and organic compounds from oxidizing fluids and biological activity over Note: Between (W/R) M of 2000 and 278 nontronite, kaolinite, quartz, pyrite, and apatite precipitated biotically, whereas nontronite, celadonite, kaolinite, goethite, talc, pyrite, and apatite abiotically.
geological time scales (e.g., Butterfield, 1990;Summons et al., 2011).Organic molecules adsorbed onto surfaces and into the interlayers of phyllosilicates include lowmolecular weight compounds, such as amino acids (Hedges & Hare, 1987) and polysaccharides (Dontsova & Bigham, 2005) as well as higher molecular weight compounds (Kennedy et al., 2002).Finally, phyllosilicates such as smectites have been proved to be able to preserve microbial filaments, benthic foraminifera, bio-generated textures, mineralized cells, and organic carbon organic cells and, for this reason, they are thought to be potential targets when searching for biosignatures of past and/or extant microbial life on Mars (Bishop et al., 2013;Vago et al., 2017;Westall et al., 2021).

Evaporation of Modeled Brines and Biosignature Preservation
Our model results indicate that evaporation of brines evolved from abiotic and biotic basalt weathering would produce different mineralogical assemblages over geological time scales (Table 4).Mainly quartz with minor Mg-clay (sepiolite) and sulfate (gypsum) would precipitate from near-neutral fluids that evaporated after bio-mediated basalt weathering, whereas Fe/Mg-clays (nontronite, talc), zeolite (analcime), and sulfate (gypsum) would form from the evaporation of alkaline brines evolved abiotically.This mineralogy is consistent with that observed in terrestrial sediments from nearneutral to alkaline saline lakes and closed basins where the rate of evaporation is higher than the influx of water from rain or tributary rivers (e.g., Bristow & Milliken, 2011;English, 2001;Pedro et al., 1978;Pozo & Calvo, 2018).Such environments, which include marginal lacustrine areas, playas, interdunal depressions, palustrine basins, and ponds developed in mudflat environments, are considered analogs to Martian aqueous systems that evolved during seasonal dry periods or in late-Hesperian when the planet started becoming dryer (e.g., Bristow & Milliken, 2011;Hurowitz et al., 2023;Kodikara et al., 2023).
Differences in mineralogical assemblages and mineral precipitation patterns in our models are controlled by thermodynamic equilibrium reactions and evolution of physicochemistry of fluids during simulations similar to those observed in natural evaporative systems.Here, change in pH, salinity, activity of water and dissolved silica, abundance of elements, and dominant anion in solution are all factors that control the reaction pathways of authigenic minerals formation during brine evaporation (Bristow & Milliken, 2011;Calvo et al., 1999;Hay & Sheppard, 2001;McHenry et al., 2023;Pozo & Calvo, 2018;Tosca, 2015).Specifically, in closed basin dominated by sulfate-and/or chloride-rich brines where dissolved Mg 2+ is present and Al 3+ is low, similar to those used in this study, the pH tends to become more alkaline as the solution becomes more concentrated via evaporation (Bristow & Milliken, 2011).This process produces chemical gradients in the aqueous environment and promotes the precipitation of authigenic minerals in specific lateral and vertical arrangements (Bristow & Milliken, 2011;Deocampo et al., 2017;Pozo & Calvo, 2018), which can be diagnostic of the evolution of local environmental conditions (Deocampo et al., 2017;Lander & Hay, 1993;McHenry et al., 2020).
Secondary minerals predicted by evaporation models in this study have been detected at various Martian sites including Gale crater, Hellas basin, and Marth Vallis (Bristow et al., 2018;Carter et al., 2013;Ehlmann et al., 2009;Hazen et al., 2023;Milliken et al., 2008;Rampe et al., 2017).Amorphous silica and sulfate deposits on the surface of Mars and in Martian meteorites are thought to have been formed following precipitation in cracks and pores during diagenesis or following evaporation of brines (Achilles et al., 2020;Bridges & Grady, 2000;Bristow et al., 2018;Changela & Bridges, 2011;Hazen et al., 2023;Milliken et al., 2008;Rampe, Blake, et al., 2020;Rapin et al., 2019;Vaniman et al., 2018), whereas Fe/Mg clays and zeolites identified by orbiters and/or rovers on Mars are usually considered secondary phases produced by the alteration of basalts in aqueous solutions at ambient and/or hydrothermal temperatures (Achilles et al., 2020;Bishop et al., 2008;Bristow et al., 2015Bristow et al., , 2018Bristow et al., , 2021;;Ehlmann et al., 2009;Kodikara et al., 2023;Milliken et al., 2010;Poulet et al., 2005;Rampe, Blake, et al., 2020;Viviano et al., 2013).However, experimental and field studies have demonstrated that Fe/Mg clays and zeolites can also form by direct precipitation or transformation of precursor minerals in near-neutral to alkaline waters at ambient temperature (Bristow & Milliken, 2011;English, 2001;Pedro et al., 1978;Tosca et al., 2011).Talc-like clays (talc, kerolite) can form at 25°C by direct precipitation from Mg-rich alkaline brines with a pH >8.5 (Bristow & Milliken, 2011;Tosca et al., 2011;Tutolo & Tosca, 2018); nontronite can form from a ferrihydrite precursor or as secondary product of the reaction of silica with Fe-oxides (Pedro et al., 1978), which are common and widespread on Mars (Du et al., 2023;Hazen et al., 2023;Rampe, Blake, et al., 2020); zeolites can crystallize from an authigenic amorphous aluminous smectite in saline Na-rich brines (English, 2001;Hay & Sheppard, 2001;Langella et al., 2001;Remy & Ferrell, 1989), a scenario that is in agreement with model results in this study that predicted zeolites forming after smectites during evaporation of abiotically evolved fluids (Figure 4).Following these arguments, we suggest that water-basalt interaction may not be the only pathway that led to Fe/Mg clays and zeolites formation in Martian basins and neo-formation of these minerals following evaporation of alkaline brines should not be entirely ruled out.
To date, Fe/Mg-smectites, zeolites, sulfates, and amorphous silica have been observed together only at a few Martian locations (Ansan et al., 2011;Ehlmann et al., 2009), with zeolites being detected only from orbit (Carter et al., 2013;Ehlmann et al., 2009;Mousis et al., 2016;Ruff, 2004;Sun & Milliken, 2015) and never reported from in situ analysis at Gale crater (Rampe, Blake, et al., 2020).Reasons that may limit or preclude the observation of zeolites on Mars include their difficult identification and distinction from polyhydrate sulfates using spectral data; their natural absence at specific locations on Mars due to short-lived period of lacustrine activity that did not provide enough time for their formation; their absence caused by transformation during chemical or physical weathering (Carter et al., 2013;Ehlmann et al., 2009;Kodikara et al., 2023;Sun & Milliken, 2015).Other than specific problems related to the identification of zeolites, other factors have been considered influential for the recognition of specific mineralogical associations on Mars, and their interpretation in the context of Martian paleoenvironments reconstructions.These include the limited possibility to unambiguously recognize sedimentary or morphological features that document lacustrine processes, the presence of detrital materials that can impede the observation of authigenic minerals, the difficulty in distinguishing the detrital or authigenic origin of specific secondary phases due to constraints of spatial resolution of onboard instruments.We note that such analysis in terrestrial samples is typically done through thin section microscopy and SEM analysis, which are not currently available on Mars.
In this context, geochemical modeling in this study provides additional information on mineralogical assemblages that may have evolved in the presence or absence of microbes in near-neutral to alkaline evaporative closed basins on Mars.This is relevant for the identification of spatial and stratigraphic patterns of secondary phases, such as clays and zeolites, for which multiple formation environments and mechanisms are possible (Bristow & Milliken, 2011;Kodikara et al., 2023).This is important also for the recognition of mineralogical associations that may inform current and future life detection missions on potential biotic processes that could have occurred in early Martian aqueous environments (Hays et al., 2017).
Amorphous silica, Mg-clays, and sulfates predicted in this study to precipitate following evaporation of a biotically evolved brine are all secondary phases that have high biosignatures preservation potential (Bishop et al., 2013;Vago et al., 2017;Westall et al., 2021).Evaporitic sulfate can trap organic matter, microorganisms, and fluid inclusions, and then preserve them for geologically relevant periods of time, providing information on environmental conditions at the dime of deposition (Panieri et al., 2010;Westall et al., 2021).Evaporitic hydrated (opaline) silica can entrap microbial filaments and lipids, microfossils, bubbles of microbially generated gases and biotically evolved fluids, and can preserve these biological materials as silicified structures and morphologies (McMahon et al., 2018;Pan et al., 2021).Similarly, Mg-clays can adsorb and preserve organic molecules and compounds into their structure (Butterfield, 1990;Summons et al., 2011;Vago et al., 2017;Westall et al., 2021).

CONCLUSION
In this study, analytical geochemistry and thermochemical modeling were combined to investigate inorganic biosignatures that can be used as evidence in the search for life in early Martian fluvio-lacustrine systems.Mineralogical and morphological changes identified during the analysis of the regolith simulant after biotic dissolution experiments are interpreted to be caused by the action of microbes that have enhanced mineral weathering and promoted secondary mineral formation and, thus, are considered potential biosignatures.The use of chemically accurate fluid and simulant regolith compositions to simulate the Rocknest environment at Gale crater increased the fidelity of the type of potential inorganic biosignatures that a selected microbial community could develop under specific Martian conditions.Following this, the results from previous studies that used a more generic Martian-like conditions are still valid and should be examined for the exact conditions they are applicable to when searching for biosignatures, for example, to settings characterized by an altered basaltic composition (Olsson-Francis et al., 2017) or more reducing conditions (Cogliati et al., 2022-iron was modeled as Fe 2+ ).All those conditions would be possible to find on Mars in ancient fluvio-lacustrine environments, enabling the potential identification of putative biosignatures by current and future in situ life detection missions or in returned samples.
Thermochemical modeling was specifically used to assist in the investigation of inorganic biosignatures that would form over geological times in the Gale crater aqueous system and in evaporitic environments associated with it.Compared to experimental simulations, model results show more significant differences between fluid chemistries and secondary alteration minerals that would evolve biotically and abiotically during waterbasalt interaction.In a biomediated fluvio-lacustrine environment, a "simpler" mineral assemblage is predicted to precipitate during long-term basalt weathering and subsequent brine evaporation.Under biotic conditions, mainly Fe-and Al-clays (nontronite, kaolinite) and quartz precipitate together in different proportions from a nearneutral solution during water-basalt reactions; during evaporation, quartz together with lower amounts of Mgclay (sepiolite) precipitate before sulfate (gypsum).In an abiotic system, Fe-, Al-, and Mg-clays (nontronite, kaolinite, talc), mica (celadonite), and Fe-oxide (goethite) precipitate in various amount from alkaline brines during basalt weathering; Fe-clay (nontronite), Mg-clay (talc), zeolite (analcime), and sulfate (gypsum) crystallize in sequence during evaporation.
The results of this study could aid the identification and characterization of inorganic geochemical biosignatures in fluvio-lacustrine systems on Mars and shed light on deposits produced during brine evaporation possibly associated with inorganic biosignatures formation in Martian lacustrine systems similar to Gale crater.The information in this study could also guide the selection of targets with the highest potential for preservation of biosignature during the ongoing and future in situ exploration missions and, in conjunction with other geological, geochemical, and biological evidence, could help to assess the presence of life in ancient Martian aqueous systems environments and their potential habitability.Our study strongly supports the necessity to use location-specific simulant compositions, fluid chemistries, and microbial communities when investigating inorganic biosignature formation over time scales relevant to Mars.Giving the complexity and variability of the geochemical processes that can occur in different water-rock systems and the different potential biological metabolisms involved, we want to stress that only by combining experimental microbiology, analytical geochemistry, and thermochemical modeling, it will be possible to fully discriminate between reaction pathways and products of abiotic and biotic processes and, thus, understand the formation of inorganic biosignatures and evaporitic minerals that would have formed in lowtemperature aqueous systems on Mars.In this context, we want to emphasize that thermochemical modeling is of great value when studying inorganic biosignature formation and mineralogical associations in a specific Martian environment because it allows to investigate reaction pathways and, thus, the evolution of fluid and rock chemistries even at low temperature when slow reaction rates make their study difficult under laboratory conditions.

FIGURE 2 .
FIGURE 2. Secondary minerals predicted from the dissolution of the regolith simulant at 25°C and 1 bar under abiotic and biotic conditions.(a)-Variation of the pH in the abiotic model; (b)-abiotic model where the pH was set as a free parameter; (c)-biotic model where the pH was set at 7.5; the diagrams show the formation of quartz, nontronite, celadonite, kaolinite, talc, zeolite, goethite, and greenalite in different proportions.Pyrite, apatite, gypsum, and halite precipitate in traces (<5 wt%) and are not represented in the diagrams.More details on the variability of the trace phases and minerals excluded during modeling can be found in supplementary file.The (W/R) D is represented by the shadowed area in between the two blue dotted vertical lines and it was calculated based on the element sodium because it is not incorporated into the microbes and into any of the expected precipitates.(Color figure can be viewed at wileyonlinelibrary.com)

FIGURE 3 .FIGURE 4 .
FIGURE 3. Fluid composition predicted from the dissolution of the regolith simulant at 25°C and 1 bar.(a)-pH variability in the abiotic model; (b)-model with pH as a free parameter that varies between 8.3 and 13.8; (c)-model with pH set at 7.5.Only elements relevant for a direct comparison between the models and the basalt dissolution experiments are shown in the diagrams (H 4 SiO 4 , Mg 2+ , Mn 2+ , K + , Ca + , Al 3+ , Na + , Fe 2+ ).Full data tables and variability of Cl À , HS À , NH 4 + , HCO 3 À , Ti(OH) 4 , HPO 4 are presented in supplementary file.The (W/R) D is represented by the shadowed area in between the blue dotted vertical lines and it was calculated based on the element potassium because it is not incorporated into the microbes and into any of the expected mineral precipitates.(Color figure can be viewed at wileyonlinelibrary.com) FIGURE 5. pH and fluid composition predicted evaporating brines taken at (W/R) M of 1000 from abiotic and biotic dissolution of the regolith simulant at 25°C and 1 bar.(a,b)evaporation model of an abiotically evolved fluid; (c,d)evaporation model of a biotically evolved fluid.Only elements relevant for a comparison with titration models in Figure 3 are shown in the diagrams (H 4 SiO 4 , Mg 2+ , Mn 2+ , K + , Ca + , Al 3+ , Na + , Fe 2+ ).Full data tables and variability of Cl À , HS À , NH 4 + , HCO 3 À , Ti(OH) 4 , HPO 4 are presented in supplementary file.(Color figure can be viewed at wileyonlinelibrary.com)

TABLE 2
. Composition of the fluids used in the abiotic and biotic titration models and in evaporation models.Ion concentration is in moles.

TABLE 3 .
Macey et al., 2023)element concentration in the experimental fluid (data fromMacey et al., 2023)and in the biotic model at (W/R) M interval of between 2000 and 278 equivalent to (W/R) D .

TABLE 4 .
List of minerals that formed in titration and evaporation models at 25°C and 1 bar.