Microbial pre-conditioning of perchlorate-bearing Martian regolith simulant: Geochemical evolution toward a plant-compatible substrate

Regolith simulant preparation, perchlorate amendment, and control materials

The regolith simulant used in this study was the Plant Trek Mars Mix (PTMM), a composite blend of three Exolith Lab simulants formulated to represent basaltic Martian terrain enriched in phyllosilicates and hydrated sulfates. The base mixture consisted of MGS-1, representing basaltic global Mars regolith, MGS-1C, representing a phyllosilicate-enriched component, and MGS-1S, representing a sulfate-enriched component. These components were combined to yield approximately 20 wt.% clay and approximately 2 wt.% gypsum, consistent with mineralogical abundances reported for Gale Crater sediments (Rampe et al., 2020).

The PTMM was sterilized by microwave pretreatment, processing 500 g batches for 10 minutes each to minimize microbial carryover from simulant handling while preserving bulk mineralogical integrity, consistent with current reproducibility recommendations for regolith-based agriculture experiments described by Fackrell et al. (2024). Subsequently, PTMM was amended with analytical-grade calcium and magnesium perchlorate salts to generate a perchlorate-bearing Martian regolith analog. The sterile PTMM + perchlorate Day 0 proxy contained 22,500 mg kg⁻¹ water-extractable perchlorate, equivalent to 22,500 μg g⁻¹ or 2.25 wt.% perchlorate. This sterile amended treatment represents the starting perchlorate condition for both the aerobic and anaerobic treatment trays before microbial pre-conditioning. This concentration is higher than many reported near-surface Martian perchlorate estimates and was therefore treated as a conservative high-perchlorate challenge condition designed to test whether microbial pre-conditioning could substantially reduce a large, bioavailable perchlorate pool prior to plant growth.

Several control and reference materials were included to distinguish perchlorate introduced by amendment from background perchlorate or process-related effects. These included an unmicrowaved PTMM-only control, which assessed background perchlorate in untreated simulant; a microwaved PTMM-only control, which assessed whether the microwave sterilization procedure introduced or altered measurable perchlorate; a sterile PTMM + perchlorate sample, which served as the amended Day 0 proxy; and a minimal-medium-only process sample, which evaluated perchlorate associated with the medium itself. During Stage I treatment, uninoculated PTMM receiving sterile mineral medium without microbial inoculum served as an additional process control for the microbial conditioning experiment. The minimal-medium sample was used only to evaluate perchlorate associated with the medium itself and was not sediment-normalized, because it was not extracted from a known mass of PTMM.

Microbial inoculum and Stage I pre-conditioning

A natural perchlorate-reducing microbial consortium was used for Stage I pre-conditioning. The consortium was derived from hypersaline sediments collected from Pilot Valley Basin in northwestern Utah, a perchlorate-bearing Mars analog environment where naturally occurring microbial communities have previously been shown to reduce perchlorate in laboratory enrichments (Lynch et al., 2019). Prior to inoculation, the consortium was enriched for 48–72 h in mineral salts medium containing 100 ppm NaClO₄, with acetate as the electron donor, to reduce lag-phase onset and stimulate perchlorate-reducing activity. Detailed source-location information and taxonomic composition of the microbial consortium are proprietary and are therefore not disclosed in this manuscript. Microbial titers and community composition were not monitored in the present geochemical study; therefore, treatment effects are interpreted from water-extractable perchlorate depletion and geochemical responses rather than from direct measurements of microbial population dynamics.

For Stage I treatment, sterilized perchlorate-amended PTMM was divided into replicate trays and saturated to field capacity with microbial inoculum suspension. Aerobic treatments were maintained under oxic incubation conditions in the laboratory, whereas anaerobic treatments were placed inside a Plas-Labs anaerobic chamber maintained with a standard N₂/CO₂/H₂ gas mixture. The anaerobic chamber atmosphere was cycled periodically to maintain oxygen-limited conditions. Aerobic trays were kept in the same room as the anaerobic chamber so that temperature, humidity, and photoperiod conditions remained as similar as possible between treatments. Treatments were maintained at 22°C under controlled humidity and a 16:8 h light:dark photoperiod. Sterile perchlorate-free mineral salts medium, without additional microbial inoculum or added NaClO₄, was added weekly to maintain a saturated matrix throughout the incubation.

Subsamples of PTMM from treatment and control materials were collected during Stage I for perchlorate monitoring and downstream geochemical analyses. Perchlorate was monitored using discrete water-extractable measurements from the amended Day 0 proxy, PTMM-only controls, and selected aerobic and anaerobic treatment time points during the 72-day Stage I incubation. Perchlorate concentrations were evaluated relative to the amended Day 0 proxy and to an EPA industrial soil screening reference (USEPA, 2026); no crop-specific perchlorate inhibition threshold was applied. After Stage I conditioning, pre-conditioned PTMM were transferred to Winston-Salem State University for Stage II plant cultivation and to Duke University for Stage I geochemical analyses. The Stage I material was not re-sterilized prior to Stage II; therefore, viable microbial cells, microbial biomass, or microbial residues may have persisted into the plant-growth phase, but microbial persistence was not directly quantified.

Sampling design and sample handling

Two sample sets were analyzed in this study. The post-Stage I sample set was collected at the end of microbial pre-conditioning and consisted of 20 total samples, including 10 aerobic and 10 anaerobic replicates from independent treatment containers. These samples were used for bulk elemental composition by ICP-MS and plant-available exchangeable nutrient analyses. A subset of 10 post-Stage I samples, including 5 aerobic and 5 anaerobic replicates, was used for pH, total carbon (TC), and total nitrogen (TN) measurements.

The post-Stage II sample set was collected 45 days after plant establishment and consisted of 10 total samples, including 5 aerobic and 5 anaerobic replicates collected from the rooting zone. These samples were used for pH, TC, and TN measurements to evaluate whether selected geochemical indicators established during Stage I persisted or changed after early plant growth.

Upon receipt at Duke University, samples retained slight residual moisture and were oven-dried at 60°C for 12 h. All samples were then homogenized, sieved to <2 mm, and stored dry prior to analysis. Sub-aliquots designated for ICP-MS analysis were further powdered to <250 μm using agate equipment to minimize contamination prior to digestion.

Water-extractable perchlorate analysis

Water-extractable perchlorate was measured using a 1:20 sediment-to-water extraction. Approximately 1 g of PTMM was placed into a 50 mL conical tube with 20 mL deionized water, shaken at 245 rpm for 30 min, and then allowed to settle overnight at refrigerated temperature. The following day, the liquid extract was collected and filtered through a 0.2 μm syringe filter prior to perchlorate analysis. Instrument-reported perchlorate concentrations were corrected for dilution and normalized to the original PTMM mass. The sterile PTMM + perchlorate treatment represents the Day 0 proxy for evaluating relative perchlorate depletion during Stage I.

Bulk elemental composition (ICP-MS; Post-Stage I, n = 20)

For total elemental concentrations of macro-, micro-, and trace metals, 0.05 g of homogenized powdered regolith was digested by microwave-assisted multi-acid digestion using an HNO□–HF–HCl acid mixture following EPA Method 3052 (USEPA, 1996). The resulting digests were evaporated to dryness, reconstituted in 2% HNO□, and analyzed by ICP-MS on a Thermo Fisher X-Series II quadrupole instrument at Duke University, using a collision/reaction cell for interference control. Calibration employed NIST-traceable multi-element standards, and QA/QC procedures included procedural blanks, duplicate digestions, and spike recoveries with target recoveries of 90–110% and analytical precision maintained at RSD < 5%.

Plant-available (exchangeable) nutrient analysis (NH₄OAc; post-Stage I, n = 20)

Exchangeable cations were extracted using 1.0 M ammonium acetate (NH₄OAc), pH 7.0. For each sample, 2.00 g of the <2 mm fraction was shaken with 40.0 mL NH₄OAc for 120 minutes, centrifuged at 3000 × g for 15 minutes, and filtered through a 0.45 μm membrane. Extracts were analyzed by ICP-OES for Ca²⁺, Mg²⁺, K⁺, Na⁺, and detectable micronutrients including Fe, Mn, Zn, and Cu. Results are reported on a dry-mass basis (mg kg⁻¹), and selected samples were analyzed in analytical replicate to verify measurement precision and reproducibility.

pH, total carbon, and total nitrogen (post-Stage I and post-Stage II)

Although the primary focus of this study is Stage I microbial pre-conditioning, Stage II samples were included to provide initial insight into how the geochemical changes induced during Stage I persisted or evolved after early plant establishment, particularly with respect to pH stabilization and biologically derived carbon and nitrogen accumulation.

Soil physicochemical properties were measured on the same two sample sets: 10 post-Stage I samples (5 aerobic and 5 anaerobic) collected after microbial pre-conditioning, and 10 post-Stage II samples (5 aerobic and 5 anaerobic) collected from the rooting zone after 45 days of plant growth.

For pH analysis, each sample was measured using a 1:1 (w/v) regolith-to-ultrapure water slurry. Subsamples of 2.00 g of the <2 mm fraction were combined with 2.0 mL ultrapure water, placed on a horizontal shaker for 30 min, and allowed to settle for 5 min at room temperature prior to measurement. The pH was determined using a calibrated glass combination electrode standardized with pH 4.00, 7.00, and 10.00 buffer solutions before each measurement. Each sample was measured in triplicate, and reported values represent the mean of the three replicate measurements.

Moreover, the same samples were finely ground for TC and TN analysis at Duke University using a high-temperature combustion elemental analyzer. Approximately 0.1 g of each sample was weighed into tin capsules and combusted at high temperature. All samples were analyzed in duplicate, with typical instrument precision of ±0.05 wt.% for carbon and ±0.005 wt.% for nitrogen. Because the PTMM simulant was formulated without carbonate-bearing mineral phases, measured TC values were interpreted as biologically derived organic carbon accumulation.

Water-extractable perchlorate during Stage I

Discrete water-extractable perchlorate measurements, reported as averages of replicate measurements when available, showed that the sterile PTMM + perchlorate Day 0 proxy contained 22,500 mg kg⁻¹ perchlorate. In contrast, unamended PTMM controls contained negligible perchlorate, with 0.0175 mg kg⁻¹ in the non-microwaved control and 0.00367 mg kg⁻¹ in the microwaved control. In the inoculated aerobic treatment, water-extractable perchlorate was 2031.7 mg kg⁻¹ by Day 3 and 1755.0 mg kg⁻¹ by Day 72, corresponding to 9.03% and 7.80% of the Day 0 proxy, respectively. In the inoculated anaerobic treatment, perchlorate was 1545.0 mg kg⁻¹ by Day 14 and 1767.3 mg kg⁻¹ by Day 72, corresponding to 6.87% and 7.85% of the Day 0 proxy, respectively (Fig. 2).

Figure 2.

Stage I bulk and plant-available elemental concentrations

Stage I bulk and plant-available elemental concentrations are shown in Fig. 4, and the complete dataset, including Welch's t-test p-values and FDR-adjusted q-values, is provided in Supplementary Table S1a,b. Each treatment included ten replicate samples (n = 10), reported as mean ± standard deviation. Bulk measurements represent the total elemental inventory of PTMM, whereas exchangeable measurements represent the operationally defined plant-available fraction. Bulk elemental compositions were broadly similar between anaerobic and aerobic treatments, indicating that microbial pre-conditioning did not substantially alter the total elemental inventory of the simulant. The main exception was phosphorus, which was significantly higher in the aerobic bulk fraction than in the anaerobic bulk fraction (2602.8 ± 112.8 vs. 1191.6 ± 69.3 mg kg⁻¹; p = 1.69 × 10⁻¹⁵, q = 2.71 × 10⁻¹⁴).

Treatment effects were more pronounced in the exchangeable fraction. Aerobic conditioning significantly increased plant-available phosphorus relative to the anaerobic treatment (222.16 ± 24.15 vs. 3.31 ± 2.47 mg kg⁻¹; p = 2.79 × 10⁻¹⁰, q = 2.79 × 10⁻⁹). Exchangeable sodium was also higher under aerobic conditions (5054.2 ± 511.4 vs. 3143.2 ± 462.2 mg kg⁻¹; p = 7.41 × 10⁻⁸, q = 2.47 × 10⁻⁷), whereas exchangeable magnesium was lower (920.4 ± 50.9 vs. 1034.9 ± 61.6 mg kg⁻¹; p = 2.50 × 10⁻⁴, q = 4.99 × 10⁻⁴). Exchangeable nickel was also lower in the aerobic treatment, although both concentrations were low in absolute terms.

pH

Soil pH values for both Stage I and Stage II are summarized in Fig. 3. Following Stage I, pH ranged from 8.76 to 8.97 in the anaerobic treatment and from 8.41 to 8.70 in the aerobic treatment. Following Stage II (45 days of plant growth), pH ranged from 8.38 to 8.57 in the anaerobic treatment and from 8.25 to 8.34 in the aerobic treatment. Thus, pH values for both treatments decreased from Stage I to Stage II, with aerobic samples showing lower pH ranges at both stages relative to anaerobic samples.

Figure 3.

Organic carbon and nitrogen content

Replicate total carbon and nitrogen concentrations for anaerobic and aerobic treatments across Stage I and Stage II are summarized in Table 1. Each condition included five replicate samples (n = 5). Total carbon remained relatively stable across treatments and stages, with no significant aerobic–anaerobic difference in Stage I (0.545 ± 0.033 vs. 0.569 ± 0.023 wt.% C; p = 0.224, q = 0.299) or Stage II (0.555 ± 0.032 vs. 0.562 ± 0.011 wt.% C; p = 0.678, q = 0.678).

Nitrogen concentrations were low, with several anaerobic replicates falling below the instrument detection limit of 0.0050 wt.%. Using one-half of the detection limit for non-detects, TN was not significantly different after FDR correction in Stage I (0.00575 ± 0.00458 vs. 0.00965 ± 0.00090 wt.% N; p = 0.130, q = 0.260) or Stage II (0.00501 ± 0.00344 vs. 0.01068 ± 0.00168 wt.% N; p = 0.0169, q = 0.0676).

Stage I microbial treatment strongly depleted water-extractable perchlorate

A central success of Stage I was the strong reduction in water-extractable perchlorate. This is important because perchlorate is one of the major chemical barriers to using Martian regolith as a plant-growth substrate. Starting from a deliberately high-perchlorate PTMM condition, microbial treatment reduced the soluble perchlorate pool by more than 90% in both aerobic and anaerobic treatments. Although the taxonomic composition of the consortium is proprietary, this response is consistent with perchlorate-respiring microbial metabolism, in which ClO₄⁻ can function as a terminal electron acceptor and is reduced through lower oxychlorine intermediates toward chloride (Kengen et al., 1999; Youngblut et al., 2016). This means that the microbial consortium likely lowered the most immediate toxicity constraint through direct perchlorate reduction, rather than only producing secondary changes in nutrient chemistry. The scale of this reduction is especially clear when compared with the EPA industrial soil screening level (USEPA, 2026). The untreated Day 0 proxy was far above this reference value, whereas the treated samples approached it much more closely by the end of Stage I. Although the final concentrations remained slightly above the screening level, the shift from a highly perchlorate-enriched starting material to a much lower residual perchlorate range demonstrates that microbial pre-conditioning can move PTMM strongly toward a less inhibitory and more plant-compatible chemical state. This provides the foundation for the later nutrient and substrate-development responses observed in the treated regolith.

pH Decline across stages indicates active biogeochemical processing

Both treatments exhibited a reduction in pH from Stage I to Stage II after 45 days of plant growth, indicating net proton-generating biogeochemical activity during cultivation. Although the substrate remained overall alkaline, this shift is important because even modest acidification can influence nutrient availability, cation exchange behavior, and trace metal mobility in basaltic and phyllosilicate-bearing materials. The decline likely reflects biologically mediated acidification driven by CO₂ hydration from microbial and root respiration, which generates carbonic acid in pore waters, as well as proton release associated with root nutrient uptake (Brady and Weil, 2008). By the onset of Stage II, the biologically derived carbon accumulated during Stage I likely provided an initial reactive carbon pool that further supported microbial respiration and rhizosphere acidification throughout the 45-day plant establishment period.

The anaerobic-treated regolith showed a slightly higher initial pH following Stage I, but the magnitude of pH decline between stages was similar in both treatments (Fig. 3), suggesting that plant-driven and microbially mediated acidification proceeded comparably regardless of pre-conditioning history. This means that the conditioned PTMM remained chemically active after plant introduction rather than behaving as an inert mineral substrate. In basaltic and phyllosilicate-bearing substrates such as PTMM, this is significant because changes in pH can alter surface protonation of Fe- and Al-(oxyhydr)oxide phases, shifting cation exchange behavior and trace metal mobility, with direct implications for nutrient accessibility (Bolan et al., 2003).

Early organic matter and nitrogen establishment during the first conditioning cycle

The regolith initially contained no deliberate organic amendment, yet measurable TC and TN were present in both post-Stage I microbial samples and post-Stage II samples after plant growth (Table 1). The persistence of these pools suggests that microbial pre-conditioning established an early organic reservoir that was maintained during initial plant establishment rather than rapidly lost from the system. This is encouraging because Stage I represents only the first conditioning cycle, yet it already produced a measurable carbon–nitrogen signal within an initially barren mineral substrate.

These results suggest that microbial biomass turnover, necromass accumulation, and extracellular polymer production during Stage I contributed to the first persistent organic matter reservoir within the mineral substrate (Buckeridge et al., 2022). This initial pool likely remained available for continued microbial processing and early rhizosphere development throughout the 45-day plant growth phase. In early-stage mineral systems, the formation and persistence of this microbially derived C–N pool represents a key phase of pedogenesis, consistent with classic observations that carbon and nitrogen accumulate during the earliest stages of soil formation (Ni et al., 2023; Wang et al., 2022).

Notably, this first-cycle organic reservoir should be interpreted as an initial threshold rather than an endpoint. The measured nitrogen pool was still small, and additional nitrogen inputs or continued biological nitrogen accumulation would likely be required for sustained crop production. However, the fact that measurable C and N were established and maintained after only one microbial conditioning cycle is encouraging. Repeated microbial conditioning, plant growth, root turnover, and harvest cycles could progressively increase organic matter retention, strengthen microbial nutrient cycling, and move PTMM further toward a stable soil-like growth substrate.

4.4 Nutrient mobilization during Stage I conditioning

The Stage I plant-available nutrient data provide the clearest indication of how microbial pre-conditioning altered the PTMM chemical environment. Phosphorus represents the most diagnostic response because aerobic conditioning strongly increased exchangeable P, whereas anaerobic samples remained near zero (Supplementary Table S1b). This contrast suggests that oxygen availability controlled whether microbial activity could mobilize phosphorus into a plant-available pool. In basaltic, Fe/Al-rich substrates, much of the indigenous phosphorus is expected to reside in surface-sorbed or partially occluded pools associated with Fe- and Al-(oxyhydr) oxide and clay surfaces, where plant availability is typically low (Johnson and Loeppert, 2006). Low-molecular-weight organic acids such as citrate, malate, and oxalate are known to mobilize phosphate from iron-rich phases through ligand exchange and ligand-enhanced dissolution (He et al., 2008). In the aerobic microcosms, sustained oxidative microbial metabolism likely promoted greater production of organic ligands, proton-generating respiration products, and mineral-surface weathering reactions, whereas the anaerobic system lacked sufficient oxidative weathering pathways to mobilize comparable phosphorus pools. Accordingly, low-molecular-weight organic ligands and other microbial weathering byproducts likely enhanced desorption of phosphate from Fe-and Al-rich mineral surfaces specifically under oxic conditions, providing a plausible mechanism for the pronounced aerobic enrichment in exchangeable phosphorus observed here.

The accompanying difference in bulk P concentrations between treatments should, however, be interpreted cautiously, because total phosphorus cannot increase without an external source. This divergence more likely reflects subsample-scale regolith heterogeneity and/or matrix effects from repeated mineral medium additions, which may leave residual salts or precipitates that inflate mass-based totals after drying.

From a plant-growth perspective, the microbial treatment appears to improve PTMM but does not make it a complete growth substrate by itself. Calcium, magnesium, and potassium were present at relatively high plant-available concentrations, suggesting that these major cations are unlikely to be the main limiting nutrients during early establishment. Aerobic treatment also produced a plant-available phosphorus pool that may be sufficient to support early growth. However, the anaerobic treatment did not show comparable P mobilization, nitrogen remained low, and micronutrients such as Zn and Cu remained limited. This means that microbial pre-conditioning can reduce some important nutrient constraints, especially under aerobic conditions, but additional nutrient management would likely still be needed for sustained plant growth and crop production.

Sodium is another important limitation. Although major nutrient supply was generally high, elevated Na is more consistent with a salinity risk than with a nutritional benefit. Excess Na can reduce water uptake and impair growth through both osmotic stress and ionic imbalance (Munns and Tester, 2008). Sodium can also disrupt K homeostasis and membrane selectivity, increasing the likelihood of physiological stress during germination and early seedling development (Volkov, 2015). Therefore, future use of microbially conditioned PTMM would likely require not only nutrient supplementation, but also salinity management through leaching, dilution, improved water control, or amendment strategies.

Micronutrient availability remained limited at this stage, particularly for Zn and Cu. In alkaline mineral substrates, cationic micronutrients commonly become less available because they precipitate or are more strongly retained at high pH (Thomas et al., 2021). This limitation does not diminish the progress achieved during Stage I, but it does highlight the importance of biological amendments during the next phase. For instance, arbuscular mycorrhizal fungi (AMF) are well known to improve the uptake of relatively immobile micronutrients, particularly in nutrient-poor and weathering-limited systems (Wahab et al., 2023). Overall, these results suggest that microbial pre-conditioning is a promising first step toward improving plant compatibility in PTMM, creating a chemically improved foundation on which later nutrient, salinity, and plant–microbe management strategies can build. Trace metal behavior adds important context to the nutrient results. Although PTMM contained detectable bulk concentrations of Cr, As, Cd, and Pb, the plant-available fractions remained at or near non-detectable levels across both treatments (Fig. 4c,d), with only a minor aerobic arsenic signal observed. Under alkaline conditions, metal mobility is often reduced because many cationic metals become less soluble, while Fe- and Al-rich mineral surfaces continue to promote adsorption and precipitation-based retention (McLean and Bledsoe, 1992). This indicates that the observed nutrient mobilization, especially for phosphorus, did not immediately come at the cost of broad toxic-metal release.

Figure 4.

Persistence and relevance of trace metals for future system design

The PTMM simulant contains chromium, nickel, cobalt, copper, lead, and arsenic at bulk concentrations consistent with a basalt-derived mineral assemblage. During Stage I, these elements remained at or near non-detectable levels in the plant-available fraction (Fig. 4c,d), indicating strong retention under the alkaline porewater conditions present during microbial conditioning.

This apparent stability should, however, be interpreted as characteristic of an early-stage system rather than a permanent geochemical state. As Stage II progresses, continued root respiration and the accumulation of rhizosphere-derived low-molecular-weight organic ligands may increase the lability of Ni, Co, Cu, and As through chelation, desorption, and ligand-promoted dissolution of Fe and Mn oxide surfaces (Qin et al., 2018). This pathway is especially relevant for arsenic, as competing organic ligands can suppress arsenate adsorption to Fe oxides and thereby enhance its mobility (Violante et al., 2010). In addition, localized acidification and transient low-redox microsites within developing aggregates may destabilize Fe(III)- and Mn-oxide phases, releasing previously sorbed Ni, Co, Cu, and particularly As.

This progression introduces an important design consideration for future closed-loop cultivation systems. The bulk concentrations of Cr, Ni, Co, Cu, Pb, and As in PTMM are sufficiently high that even modest increases in their bioavailable pools during later growth cycles could become relevant for both plant performance and food safety. Nickel and chromium may interfere with essential Fe and Mg uptake pathways, whereas Pb and As remain of particular concern because of their potential translocation into edible biomass (Kabata-Pendias, 2011). Stage II incorporates an inoculum of multispecies AMF, which may help moderate these risks by stabilizing mineral-associated organic matter, increasing sorption surface complexity, and reducing net metal bioavailability through hyphal sequestration and rhizosphere buffering (Dhalaria et al., 2020). However, the extent to which AMF can suppress trace metal mobilization in a perchlorate-modified basaltic simulant remains uncertain and should be explicitly monitored in future multi-cycle experiments.