A comparison between polyacrylamide and arcillite granules as substrates for space agriculture

Plant cultivation in space requires substrates that provide sufficient water retention, root system support, and nutrients. Arcillite – a calcined montmorillonite clay product – is currently used in growth chambers on the International Space Station despite significant variability in mineral composition and occasional detrimental effects on seed germination and plant growth. In contrast, hydrogels – which are used as soil amendments in agriculture – can be produced with desired structural and chemical properties. We compared two batches of arcillite and polyacrylamide (PAAM) hydrogel granules by growing and analyzing radishes (Raphanus sativus) and observed significant differences in leaf and bulb biomass, evapotranspiration, and photosystem II efficiency between the two batches of arcillite, with soluble calcium, magnesium, selenium, and sulfur being more abundant in the poorly performing arcillite. There was no difference in leaf and bulb biomass or photosystem II efficiency between PAAM and the arcillite batch without excessive calcium, but radishes grown in PAAM required significantly less water. Our results indicate that PAAM can outperform arcillite in water-limited environments because of reduced evapotranspiration rates, greater uniformity, and comparable yields. Therefore, hydrogels such as PAAM may be beneficial for space plant cultivation.

As space agencies prepare for long duration missions to the Moon and Mars, developing a sustainable fresh food supply to supplement crew diets and identifying pathways to recycle resources have become priority objectives. Central to self-sustained space travel without the need for frequent resupply missions is the development of bioregenerative life support systems which use crops to convert biological waste (graywater, solids, and CO 2) into fresh food and clean water (Ferl et al., 2002; Wheeler, 2011). Decades of technological development has resulted in improved growth chambers like the Advanced Plant Habitat (APH) for testing cultivation methods, candidate crops, and investigating how higher plants respond to spaceflight-specific conditions such as variable gravity, increased ionizing radiation, and atypical atmospheric compositions (Campbell et al., 2001; Halstead and Dutcher, 1984; Hasenstein et al., 2022; Kiss et al., 2019; Porterfield et al., 2003; Zabel et al., 2016).

In the APH, water is supplied through porous tubes which are surrounded by arcillite to facilitate water distribution throughout the root zone (Monje et al., 2020). While environmental factors (e.g., temperature, humidity, CO 2 concentrations) are largely controlled in the APH, reduced gravity and increased importance of surface interactions and viscosity result in modified water gradients (Jones and Or, 1999). To maintain root oxygenation in microgravity, arcillite is selected with a particle size between 1 and 2 mm to simultaneously sustain air- and water-filled pores, and to prevent root hypoxia (Jones et al., 2012; Jones et al., 2002; Steinberg et al., 2002).

However, arcillite is known to have a high variance in its intrinsic nutrient content. Without proper conditioning, changes in soluble minerals beyond the tolerance level of specific crops will lead to stunted plant growth (Adams et al., 2014; Gambrel et al., 2010; Spomer et al., 1997; Warren and Bilderback, 1992). Conditioning typically includes washing arcillite to reduce the concentration of excess or potentially toxic solutes followed by either the addition of slow-release fertilizer prills or infusion with established hydroponic nutrient solutions. Additionally, the rigidity of arcillite and shallow depths of APH root modules may be detrimental for the expansion of plant structures such as radish bulbs (John et al., 2021; Massa et al., 2017).

Terrestrial agriculture uses hydrogel soil amendments to improve water retention and reduce the compaction of clay-heavy soils. In regions with arid climates or poor-quality soils, hydrogels have been shown to increase crop yield while reducing irrigation requirements (Akelah, 1990; Landis and Haase, 2012; Rudzinski et al., 2002) and, as manufactured materials, they have the advantage of well-defined and customizable chemical, physical, and hydraulic properties (Bansil and Gupta, 1980; Baselga et al., 1987; Bhadani and Mitra, 2016; Caulfield et al., 2002). Hydrogels can retain water and slowly release agrochemicals (e.g., fertilizers and biocides) and therefore may also be beneficial for crop production in space. Since hydrogels can be packaged dry and rehydrated as needed, they markedly reduce system mass and storage volume relative to arcillite. The ion exchange capabilities of some hydrogels could be advantageous for removing contaminants and heavy metals from regolith (Oze et al., 2021; Xie et al., 2010). Thus, hydrogels are now being investigated as growth substrates in microgravity and future lunar or Martian outposts (Hasenstein and Miklave, 2024; Peyrusson, 2021; Teng et al., 2022).

While the benefits of hydrogel soil amendments are well documented, it is not evident if these effects persist when hydrogels are used as the sole substrate. While ionic hydrogels such as sodium polyacrylate lead to salt stress and a reduction in crop yield, nonionic hydrogels like polyacrylamide should not introduce ion loads. We compared the biomass and photosynthetic efficiency of radish plants (Raphanus sativus) grown in polyacrylamide granules against two distinct batches of arcillite. Improved water use efficiency and equivalent yields, combined with expected benefits to system mass and storage volume, indicate that granular polyacrylamide is a promising substrate for space agriculture.

Substrate preparation

A solution containing acrylamide (96 g/L), N’N’-methylenebisacrylamide (4.0 g/L), and ammonium persulfate (4.0 g/L) was dispensed as 80 µL boluses into 85 °C silicone oil (Silicone Solutions, Cuyahoga Falls, OH, United States) to induce polymerization. The resulting polyacrylamide granules (PAAM) were rinsed with deionized water and stirred overnight to remove silicone oil and residual monomers and then dried at 50 °C. Dry PAAM was stored at room temperature in glass bottles.

Two batches of arcillite (Turface Pro League Elite) were acquired in 2019 and 2024 (hereafter referred to as ARC19 and ARC24, respectively) from Turface Athletics (Buffalo Grove, IL, United States) and originally intended for flight experiments. Each batch was sifted to obtain the 1–2 mm size fraction (Jones et al., 2012; Steinberg et al., 2002), rinsed with deionized water until free of fine particulates, and then dried at 50 °C.

Containers modeled after APH science carrier root modules (John et al., 2021; Monje et al., 2020) were filled with either 1,300 g of ARC19 or ARC24, or 100 g of dried PAAM. Each container received 1,500 mL of a chloride-free half-strength Murashige and Skoog (MS) nutrient solution (John et al., 2021) and was left overnight for equilibration. During the rehydration of PAAM, a 1 mm mesh screen was placed between the substrate and container lid to retain granules. Two root modules were prepared for each substrate.

Crop selection

We used Raphanus sativus var. “Cherry Belle” seeds (Mountain Valley Seed, Salt Lake City, UT, United States) because of their rapid growth, low canopies, and extensive history as a target for space biology research (Fritsche et al., 2024; Hasenstein et al., 2025; John et al., 2021; Stutte et al., 2004; Wehkamp et al., 2012).

Planting and cultivation

Floral foam disks, 1 cm thick and 5 cm in diameter, were cut from blocks of Oasis Aeromax foam (Smither-Oasis, Kent, OH, United States) and four radish seeds were inserted 5 mm deep into each disk. Five foam disks were then installed into the lids of each substrate-filled root module. Additional deionized water was added to the containers until the surfaces of the foam disks were saturated. Each assembled container was weighed and then wrapped with aluminum foil and placed in an environmental chamber (Percival Scientific, Perry, IA, United States). Three days after planting, the foil was removed and 514–529 μmol m −2 s −1 of photosynthetic photon flux density (PPFD, Supplementary File 3) was provided with a 16/8-h light/dark cycle. Light spectra were measured with an ILT350 Chroma Meter (International Light Technologies, Peabody, MA, United States) Environmental conditions fluctuated from 24 °C during the day period and 20 °C during the night, and CO 2 between approximately 450 and 750 ppm.

Three days after planting (DAP), and every subsequent day thereafter, water loss was measured gravimetrically. Seven DAP, DI water was added daily to replenish what was lost through evapotranspiration in the previous 24 h. Seedlings were thinned 10 DAP to the healthiest individual in each floral foam disk. The remaining plants were harvested on DAP 28.

After 28 days, the substrates were thoroughly washed with DI water to remove root fragments and remaining nutrients, and then autoclaved and dried. Small quantities of PAAM (approximately 500 mg) were removed and stored for assessing changes in water holding capacity and chemical structure. Arcillite and the remaining PAAM were then re-infused with 1,500 mL of half-strength MS nutrient solution for the next grow-out. This infusion, cultivation, and harvest process was repeated for three sequential grow outs.

Radish harvest

Before harvesting, we measured photosystem II efficiency (Φ PSII) as the quantum yield of chlorophyll from five randomly selected locations on mature and young leaves of each plant with a Fluorpen FP 100 (Photon Systems Instruments, Drasov, Czech Republic). The fresh weight of leaves was measured immediately after their removal from each plant. Leaves and were photographed for area measurements with FIJI/ImageJ (Schindelin et al., 2012). The fresh weight of radish bulbs was taken after removing fragments of floral foam and substrate. All leaves and bulbs were dried at 60 °C for 7 days before determining the dry weight.

Nutrient and mineral analysis

Radish seeds and dried leaf and bulb tissue of each plant were pulverized by mortar and pestle, with up to 150 mg of the powdered tissue precisely weighed into 15 mL centrifuge tubes. To each of these, 2 mL of trace-metal-grade nitric acid (Fisher Scientific, Pittsburgh, PA, United States) was added. Sample tubes were thoroughly shaken and tissue digested for 3 days before being diluted with 10 mL of NanoPure (20 MΩ) water. The aqueous extract was transferred into clean centrifuge tubes through Zenpore Qualitative Grade 4 filter paper (Fluted Filter, Fo Tan, N.T. HKG).

Substrate samples were prepared in triplicate from unused substrate that had been washed and dried to remove silicone oil (PAAM) or fine particulates (arcillite). 15 g of dry arcillite or 1.15 g of dry PAAM was combined with 30 mL of NanoPure water and thoroughly shaken and left to extract for 24 h. Three mL aliquots were passed through Millex GV 0.22-micron syringe filters (Sigma-Aldrich, St. Louis, MO, United States) and diluted with 7 mL NanoPure water and 2 mL nitric acid. Samples for PAAM monomers were prepared by dissolving up to 150 mg of reagent (i.e., acrylamide, N’N’-methylenebisacrylamide, or ammonium persulfate) in 10 mL NanoPure water and adding 2 mL nitric acid.

The mineral content of all samples was measured with inductively coupled plasma optical emission spectroscopy (PerkinElmer Optima 5300 DV). ICP emission intensities were calibrated using standards with ion concentrations between 0 and 50 mg/L with Yttrium as an internal standard. Quality control checks were performed after every 15 samples by reanalyzing intermediate-concentration calibration standards. Calibration curves resulted in R 2 values greater than 0.99.

Polyacrylamide water holding capacity

Two sets of dry, unused, PAAM – approximately 100 mg each – were photographed with a scale bar after hydrating in either 50 mL deionized water or 50 mL of MS media, and again after drying at 60 °C. This was repeated for PAAM samples taken before each subsequent grow out. The images were analyzed with FIJI/ImageJ (Schindelin et al., 2012) to measure the major and minor axes of best-fit ellipses. The swelling ratio of PAAM was calculated as the ratio of hydrated major axis divided by the dry major axis.

Statistical analysis

The effect of substrate type and reuse on the biomass of radish leaves and bulbs, water use, leaf mass per area, photosystem II efficiency, and tissue nutrient contents were examined by two-way ANOVA between substrate type and reuse, with interactions, and a significance threshold of α = 0.05. When substrate reuse was not found to be a significant or interacting factor, the datasets from each grow out were pooled and one-way Tukey-Kramer post hoc tests were performed with substrate type as the explanatory variable. Otherwise, two-way Tukey-Kramer post hoc tests were performed. All post hoc tests used a family-wise error rate (FWER) of α = 0.05. Variables that were not significantly different with substrate type or reuse were excluded from subsequent analyses. ANOVA tests were performed on the level of individual plants for all data except water use efficiency,

The concentration of nutrients in leaves and bulbs was normalized (μ = 0, σ = 1), and a principal component analysis (PCA) was performed to construct linearly independent parameters. Only principal components with eigenvalues greater than 1 were retained for correlation analysis. Correlations between principal components and yield, water use, and leaf functional traits were calculated as the Kendall τ B correlation coefficient, and the significance of these correlations were determined via τ-test based on a FWER of α = 0.05 with Holm-Bonferroni corrections.

Evapotranspiration and fresh biomass

Daily evapotranspiration rates did not significantly differ between arcillite and polymer substrates until DAP 12 (Figure 1A). Thinning seedlings on DAP 10 did not decrease evapotranspiration during the subsequent 24 h, suggesting that plant-related water loss is minimal in this initial phase of growth and water loss is instead dominated by substrate evaporation. As evaporation from the substrate is expected to remain constant, changes in the rate of water loss during the second half of the experiments (DAP 15–28) can be attributed to transpiration. Plants grown in ARC24 had constant rates of water loss and consumed a total of 2.1 ± 0.6 L per container, significantly less than ARC19 (4.6 ± 0.3 L) and PAAM (3.5 ± 0.6 L). Unlike ARC24, the rate of water loss for PAAM and ARC19 increased over time, and ARC19-grown plants had significantly higher total evapotranspiration than PAAM (Figure 1A). Since water was replenished by weighing planting containers, rather than by measuring soil moisture content, the actual water supply decreased as radish biomass increased. However, the largest yield per container (i.e., the combined leaf and bulb fresh weight of all 5 plants) was 165 g, and the actual water supply decreased by less than 10% over the growth period.

The increase in evapotranspiration rates for ARC19 and PAAM started around DAP 12 and coincided with the development of the first pair of true leaves. While plants growing in ARC24 also developed their first true leaves at this time, subsequent vegetative growth was reduced (Figure 1B). Radishes grown in ARC24 produced less leaf tissue than ARC19 and PAAM, whereas no difference was seen in leaf biomass between ARC19 and PAAM. Bulbs from ARC24 were also significantly smaller than ARC19, but no differences in bulb mass were found between either batch of arcillite and PAAM. These patterns were consistent for each substrate and across subsequent grow outs, as repeated use of substrates had no effect on cumulative evapotranspiration or fresh biomass.

Photosystem II efficiency

The quantum yield of photosystem II (Φ PSII) was significantly affected by the type of substrate, and significant interactions were found between substrate type and the age of leaves – either mature (fully expanded) or young (still expanding) (Figure 2). In PAAM and ARC19, Φ PSII was lower in mature leaves than young leaves (by 3.8% and 3.3%, respectively), but ARC24-grown radishes showed no change in Φ PSII between mature and young leaves. Furthermore, in young leaves the photosystem II efficiency was uniform between all substrates. No differences were seen in Φ PSII between PAAM and ARC19, regardless of leaf age.

Leaf mass per area

Leaf mass per area (LMA) was calculated by dividing dried leaf mass by leaf area, for each plant. LMA in the first grow out was significantly higher for plants grown in PAAM, and no differences were found for ARC19 or ARC24 (Figure 3).

Polyacrylamide water holding capacity

The major axis of unused PAAM granules swelled about 2.5-fold when hydrated with deionized (DI) water or MS medium (Figure 4). After the first grow out PAAM granules absorbed significantly more DI water, but the water holding capacity did not change after additional use. No change in PAAM water holding capacity was seen for PAAM when hydrated with half-strength MS nutrient solution.

Radish nutrient content

The dry weight ion content of radish tissue was significantly affected by substrate. Macro and micronutrients (Figure 5; green and red frames respectively) were found within expected ranges and beneficial nutrients (Figure 5; blue frames) showed varying concentrations. Except for manganese, the dry-weight concentrations of macro- and micronutrients in leaf and bulb tissue were within an order of magnitude of seeds.

Macronutrients

On average, sulfur was the most abundant nutrient in dried leaves and bulbs, and S content was significantly higher for ARC19 and ARC24 than for PAAM. The highest leaf concentrations of potassium were found in ARC19-grown radishes, and there were no differences between PAAM and ARC24 or for radish bulbs grown in any substrate. Calcium was significantly more abundant in leaves and bulbs of ARC24 radishes and was more abundant than K in ARC24 leaves. Leaf magnesium content was highest in ARC19 and lowest in PAAM, while ARC19 and ARC24 had similar bulb Mg. Phosphorous was the only nutrient that was not affected by substrate.

Micronutrients

The leaf contents of manganese and iron were elevated in ARC19, but no differences were found for bulb Fe. Conversely, molybdenum abundance was lower in ARC19 relative to ARC24 in leaves and bulbs, but lower than PAAM in leaves. Zinc was the only nutrient which was more abundant in leaf tissue grown in PAAM than from either arcillite. ARC24-grown radishes showed elevated copper content than PAAM-grown plants, and nickel was similarly elevated in bulb tissue.

Beneficial nutrients

Of the beneficial nutrients (Marschner, 2012) detected in radish tissues, silicon and aluminum showed no differences in bulb concentration regardless of substrate. Selenium, though, was elevated in ARC24 bulbs relative to PAAM and leaf Se was higher in ARC24 than PAAM or ARC19 grown plants. Leaf Si for ARC24 and leaf Al for PAAM were both lower than ARC19.

Substrate nutrient availability

The total nutrient availability of each unused substrate (soluble at neutral pH), in addition to the added supply from the half-strength MS nutrient solution, is shown in Figure 6. While identical amounts of nutrient solutions were added to each substrate-filled container, the nutrient availability of each substrate led to higher quantities of S, Ca, Mg, and Se in ARC24 than ARC19 or PAAM. Differences in nutrient availability were minimal between ARC19 and PAAM substrates. PAAM granules contained Si, Cu, Mo, Ni, and Zn although these ions were not detected in stock reagents (i.e., acrylamide, bis-acrylamide, and ammonium persulfate), they may be present in the silicone oil used to induce polymerization (Figure 7; Supplementary file 2).

Aluminum was not detected in any substrate at neutral pH, nor in seed stock, and is not contained in the nutrient solution – despite showing up in leaf and bulb tissue (Figure 5). No substrate was a notable source of K, P, or Mn.

Principal component and correlation analysis

A principal component analysis (PCA) was performed using the leaf and bulb composition of all nutrients that varied significantly by substrate (Figure 5). Six principal components were obtained with eigenvalues greater than 1, which account for 83.95% of the total variance of the plant ionome (Figure 7; Supplementary File 4). Principal components 1, 2, and 3 discriminate between substrate types (Figure 7B; Supplementary File 3) and significantly correlate with substrate nutrient availability (Figure 7C). However, PC 1 and 3 do not correlate with biomass or water use efficiency. PCs 4 and 5 correlate the plant ionome with water use efficiency and mature leaf Φ PSII, respectively, but not with substrate nutrient availability.

Radishes growing in polyacrylamide granules have increased water use efficiency

The ratio of fresh leaf and bulb biomass to cumulative water consumption, or water use efficiency (WUE), represents the amount of net carbon fixation for a utilized quantity of water (Briggs and Shantz, 1913). Radishes growing in PAAM substrate had higher WUE than those growing in either batch of arcillite (Figure 1). This increase in WUE is consistent with other studies which showed the benefits of hydrogels as soil amendments for bent grass (Agrostis stolonifera) and corn (Zea mays). The magnitude of these benefits were dependent on soil texture, as coarser soils showed a greater increase in WUE (Agaba et al., 2011; Agaba et al., 2010; Albalasmeh et al., 2022; Buchmann et al., 2025; Mohawesh and Durner, 2019; Terry and Nelson, 1986). Increases in crop yield and WUE from hydrogel amendments are also seen in integrated agricultural systems such as soybean-wheat intercropping, aerobic rice cultivation, and wastewater-irrigated spinach (Dhiman et al., 2021; Rajanna et al., 2022; Rehman et al., 2011). However, despite improving WUE in radishes, the leaf and bulb biomasses showed no difference between ARC19 and PAAM (Figure 1B).

Increases in WUE by hydrogels in field agriculture are broadly linked to reduced runoff and percolation. However, closed containers like APH-style root modules do not lose water by percolation and are improving WUE independent of water retention. An increase in WUE for hydrogel-amended soils can be associated with modifications to the soil-water characteristic and decreasing soil matric potential (Albalasmeh et al., 2022; Baasiri et al., 1986; Mohawesh and Durner, 2019). As the measured evapotranspiration rate during the first 12 days was identical, increased WUE cannot be related to differences in surface evaporation (Figure 1A) and may instead be linked to changes in substrate matric potential.

Photosystem quantum yield indicates substrate-specific stress

Photosystem II quantum efficiency (Φ PSII) represents the fraction of absorbed light energy that is used for carbon fixation, and is a common tool for quantifying environmental stress (Santos et al., 2022; Greer and Halligan, 2001; Lichtenthaler, 1988; Wang et al., 2022). Φ PSII increases as photosynthetic systems develop before reaching a steady-state at leaf maturity, before declining during senescence (Bielczynski et al., 2017; Hikosaka et al., 1994; Kitajima et al., 2002; Nath et al., 2013). Excess light energy can produce reactive oxygen species that elicit stress responses and reduce photosynthetic efficiency but can be quenched by increasing antenna pigments, light harvesting complexes, and electron transport chains in chloroplast (Asada, 2006; Ciesielska et al., 2025). Although all plants developed their first true leaves around 12 DAP, the leaf biomass of plants grown in ARC24 was significantly reduced (Figure 1B). Biomass accumulation or growth relies on CO 2 fixation, nutrient availability, and sufficient light energy for primary metabolism, while imbalances in resource availability reduce growth.

For ARC19 and PAAM, expanding (young) leaves had quantum yields which were greater than that of fully developed (mature) leaves (Figure 2). This decrease in quantum yield from young to mature leaves was not seen in ARC24 and correlates with an overall reduction in growth rates (Figure 1B). Correlations between leaf age and Φ PSII can depend on nitrogen availability, and the lack of a decrease in Φ PSII with leaf age suggests a stress response (de Souza et al., 2025; Field and Mooney, 1983; Niinemets, 2016), as drought, salinity, or nutrient toxicity can decrease photosynthetic efficiency (Dong et al., 2025; Kökten and Nowicka, 2025; Liu et al., 2025; Okoń and Nosalewicz, 2025; Sperdouli and Moustakas, 2012). Although ARC19 and PAAM showed higher transpiration rates and a decrease in Φ PSII of older leaves, this response is not related to drought stress, as the daily water replenishment was less than half of the available supply (Figures 1, 2).

Radish LMA suggests polyacrylamide nitrogen is bioavailable

Leaf thickness and cell size affect the total amount of light energy absorbed and can serve as a time-integrated record of the environmental conditions a plant has experienced. Leaf mass per area (LMA) is a generalized metric of thickness, regardless of leaf shape (Aneja et al., 2025; John et al., 2017; Pearce et al., 1968; Poorter et al., 2009). LMA strongly correlates with leaf nitrogen content and photosynthetic capacity, and there is evidence that elevated CO 2 further magnifies nitrogen-related increases to LMA and growth rate (Knops and Reinhart, 2000; Niinemets, 2016; Rosati et al., 2000). The standardized light conditions of our experiments imply that any changes in LMA are substrate related. No significant variation was seen among substrate types and the range of LMA (between 30–100 g/m 2 (Figure 3) was consistent with values typical for herbaceous dicots (Poorter et al., 2009). Although all substrates were washed, sterilized, and infused with the same amount of nutrient solution at the start of each experiment, the data suggest that reusing substrates leads to decreased LMA. This observation was most prominent for PAAM and may be linked to the reduced availability of nitrogen.

The swelling behavior of hydrogels is consistent over repeated cycles of hydration and drying and the maximum water content of polyacrylamide is not generally dependent on the dissolved ion activity (Bhadani and Mitra, 2016). However, for ionic hydrogels (e.g., polyacrylic acid) the water holding capacity does depend on the ionic strength of absorbed solution (Okay and Sariisik, 2000). Under acidic conditions the amide side chains of polyacrylamide can be liberated via hydrolyzation, converting up to half of the polymer’s acrylamide units into acrylic acid and releasing ammonium (Caulfield et al., 2002; Kalaleh et al., 2015; Yasuda et al., 1988). After the first grow out the water holding capacity of PAAM granules increased but remained at their original size when reinfused with MS nutrient solution (Figure 4). It is possible that the release of acidic root exudates hydrolyzed the PAAM, which would have liberated ammonium as a readily available nitrogen source for plants and microbes (Gaytán et al., 2021; Joshi and Abed, 2017; Kay-Shoemake et al., 1998). As no further changes in water holding capacity were seen after subsequent grow outs, we suspect that any available PAAM-derived ammonium was released during the first grow out and may explain the observed changes in LMA upon substrate reuse (Figure 3).

Ionomic profiles reveal substrate-specific effects

Tissue concentrations of ions (the ionome) were significantly affected by substrate, except phosphorus (Figure 5). Since most functions in plants require many nutrients, attributing growth rates or significant differences in plant traits to the availability of individual nutrients (Figure 6) is not possible (Baxter, 2015; Reid and Hayes, 2003), and is further complicated by the active regulation of nutrient uptake, plant-specific requirements, and the synergistic and antagonistic interactions between substrate and uptake processes (Jakobsen, 1993; Mengel et al., 2001; Roques et al., 2013). Instead, PCA can identify correlating nutrient interactions.

The elevated concentrations of Mg, Mn, Fe, and K in leaves significantly correlated with cumulative evapotranspiration (Figure 7). Mg, Mn, and Fe are key ions in the oxygen evolution complex that drives redox reactions, while K + signals are important for stomatal control, transpiration, and control of osmotic pressure (Mengel et al., 2001; Nieves-Cordones et al., 2016; Uchida, 2000; Zhang et al., 2022). In K-deficient substrates, transpiration rates and stomatal conductance are reduced (Brag, 1972; Sperling et al., 2024). However, while the K content of radish tissue was significantly different among ARC19, ARC24, and PAAM, the K availability was determined by the provided MS medium (Figure 6). Substrate Mn and Fe concentrations were similar between all substrate types, but these ions differed in leaf tissue from each substrate (Figure 5). Despite ARC24 containing significantly more soluble Mg than ARC19 or PAAM, Mg was highest in leaves of radishes grown in ARC19. Assimilation rates of K, Mg, Mn, and Fe may have been antagonized by the high calcium content of ARC24 (Han et al., 2019; Jakobsen, 1993; Maas et al., 1969; Rietra et al., 2017; Roques et al., 2013; Tang and Luan, 2017) but there was no direct correlation between substrate Ca and leaf tissue concentrations of these ions (Figure 7). PC 2 showed a negative correlation with substrate Cu availability which, like Ca, was more abundant in ARC24 and could have antagonized the assimilation of Mn and Fe (Rai et al., 2021).

The variance in leaf Ca content is captured primarily by PCs 1 and 4, which was associated with decreased water use efficiency (Figure 7C). The elevated Ca content in ARC24 and the reduced plant biomass and evapotranspiration rate is consistent with studies that link excess calcium to reduced growth and photosynthetic efficiency (Guo et al., 2023; Weng et al., 2022).

PC 2 also negatively correlates with the substrate availability of Selenium. This element was detected in all tissue samples, despite not being supplied by the nutrient solution and only contained in ARC24. Se can significantly affect the uptake of cations, dependent on concentration and oxidation state (e.g., selenite, SeO 3 2− or selenate, SeO 4 2−). At the concentrations found in ARC24 (∼1 mg/kg), selenate increases cation absorption while selenite decreases cation absorption (Longchamp et al., 2016). While no data is available on the form of selenium that is present in ARC24, the negative association between Se availability and tissue concentrations of many cations (Figure 7) suggests selenite as the prevalent form. The dependency of cation absorption on the specific form of low-abundance (sub-ppm) nutrients highlights the challenges of using materials with an undefined and highly variable chemical structure, and our ionomic analysis can identify ions in substrate that may affect nutrient uptake and concentration in plants.

Prospects for space applications

In this study, all plants were grown concurrently in a shared growth chamber to standardize the effects of atmospheric and lighting conditions (temperature, photon flux, partial CO 2). Differences in assessed plant traits such as photosystem II efficiency are therefore associated with the substrate. Space-specific environmental factors – like microgravity and increased ionizing radiation – may not reduce photosystem health on their own but have been shown to exacerbate the detrimental effects of excessive light and CO 2 (Karthiksaran et al., 2026; Stutte et al., 2005). As a result, substrates that cause stress responses in ground studies, as exemplified by the differences between ARC24 and ARC19, are not appropriate for space cultivation.

Food production and environmental control for long-term space expeditions and Lunar/Martian habitats will likely rely on plant production with minimal substrate volumes (Monje et al., 2003; Paul et al., 2022; Wamelink et al., 2014). Hydrogel-based substrates will benefit space cultivation by their reducing mass and dry storage volume. Because hydrogels like PAAM are synthesized, they combine stability and customizability for specific crops, nutrient requirements, and copolymerization or infusion with other organic compounds that improve crop cultivation like growth regulators, vitamins, and amino acids (Yadav and Anand, 2026).

Space agriculture and exploration can further benefit from secondary hydrogel applications to improve in-situ resource utilization. For instance, while the high concentration of perchlorates in Martian regolith severely inhibits plant germination and growth, chitosan-based hydrogels are already used to remove perchlorates from drinking water (Xie et al., 2010). Our data shows that PAAM can be beneficial for crop production during extended spaceflight missions (e.g., during transit from Earth to Mars), and studies suggest that spent substrate can then be repurposed to condition regolith and provide an organic matter resource for microbes to initiate soil formation (Grula et al., 1994; Oze et al., 2021; Peyrusson, 2021; Sutter et al., 2017). Alternatively, extremely hygroscopic hydrogels (Al-Handawi et al., 2023; Nandakumar et al., 2018) could be designed which maintain water potentials at the root zone by scavenging humidity, reducing the need for condensators for water management in microgravity (Hasenstein and Miklave, 2024). Although our results demonstrate that PAAM granules represent suitable low-mass substrates which reduce water requirements while maintaining healthy plant growth, their application for space agriculture is subject to continued investigations on long-term reusability and altered hydraulic behavior in reduced gravity. While increased hydrogel application rates can improve functions like erosion resistance and water regulation, those benefits may come at the expense of ecological function (Xu et al., 2025). Therefore, hybrid substrates of e.g., arcillite or regolith and hydrogel granules may provide benefits for space agriculture that pure hydrogel would not achieve. Furthermore, chemical or structural changes of PAAM due to hydrolysis or from many wetting-drying cycles in extended growth cycles can compromise durability and may lead to increased fragmentation; the identity, accumulation, and transportation of degradation products in plant tissues, closed-loop water systems, and microbial bioreactors need to be well-described; and best practices for substrate sterilization (e.g., UV, heat, or chemical sanitization) are unknown.