The experiments used surface soil collected from a paddy field in Ganzhou City, Jiangxi Province, China, where the total arsenic concentration was 36.9 milligrams per kilogram. That exceeded the Chinese risk-screening value of 30 milligrams per kilogram for paddy soils with a pH of 5.5 or lower. Researchers applied microscale zero-valent iron, or mZVI, at 1% by dry soil weight in both the microcosm and pot treatments, while untreated soil received the same fertilizer as a control. The microcosms were incubated anaerobically for 70 days and then aerobically, and the pots supported indica rice through a complete growing season in 2025. Each treatment had three replicates, limiting statistical power and making the findings evidence from a controlled experiment rather than a demonstration of performance across field conditions.

Arsenic release into water fell sharply after mZVI was added. On day 36 of anaerobic incubation, dissolved arsenic in the overlying water averaged 200.9 micrograms per liter in controls but only 12.5 micrograms per liter in amended microcosms, a reduction of 93.8%. At day 70, the respective concentrations were 93.6 and 8.89 micrograms per liter, corresponding to a 90.5% reduction. The difference persisted after oxygen exposure: on day 106, dissolved arsenic remained 29.9% lower with mZVI. Sequential extraction also showed that the treatment reduced the two most labile, operationally defined arsenic pools. Non-specifically adsorbed arsenic fell from 5.89% to 0.47% of total arsenic after anaerobic incubation, while specifically adsorbed arsenic declined from 13.2% to 8.63%. Following aerobic incubation, the corresponding values remained lower with mZVI: 0.24% for non-specifically adsorbed arsenic and 8.47% for specifically adsorbed arsenic, compared with 1.10% and 14.48% in controls.

The rice experiment extended the reduction in arsenic mobility to plant uptake. During cultivation, porewater arsenic remained below 10 micrograms per liter in mZVI-treated pots, whereas controls reached a maximum of 135 micrograms per liter. Iron remained elevated in amended porewater, rising as high as 64.5 milligrams per liter at day 27 and remaining at 28.9 to 38.2 milligrams per liter later in the season. MZVI increased porewater and overlying-water pH by 0.17–0.53 and 0.08–0.51 units, respectively, relative to controls, and maintained more negative oxidation-reduction potential throughout the rice-growing period. Its largest effect on that potential was a 367-millivolt decrease at day 48. These changes were consistent with active iron corrosion and the formation of secondary iron minerals that can retain arsenic through adsorption, coprecipitation and surface complexation. Dissolved organic carbon was also generally lower in amended porewater, by 50.8–105 milligrams per liter during incubation, although the researchers did not resolve how this difference contributed to methane production.

Arsenic concentrations decreased in every major rice tissue, with the greatest reductions in the above-ground portions and a smaller but still substantial reduction in edible grain. Compared with controls, mZVI lowered arsenic in roots, stems, leaves and grains by 83.3%, 92.9%, 87.4% and 34.0%, respectively. In brown rice from controls, inorganic arsenic accounted for 72.8% of total arsenic, with arsenic(V), dimethylarsinic acid and arsenic(III) accounting for 54.3%, 27.2% and 18.5%, respectively. MZVI reduced grain arsenic(V) by 68.2% and dimethylarsinic acid by 77.3%. Grain yield did not differ significantly, although rice biomass increased. The authors therefore found no evidence in this greenhouse trial that arsenic remediation required a sacrifice in grain production, but they did not quantify how much of the biomass increase arose from reduced arsenic toxicity, improved iron nutrition or a more favorable pH.

Direct analysis of the iron plaque on rice roots complicates a simple explanation based on greater surface retention. Dithionite-citrate-bicarbonate extraction showed that plaque iron increased from 17.8 to 52.7 grams per kilogram, but plaque-associated arsenic fell from 188 to 33.6 milligrams per kilogram. This pattern is consistent with mZVI restricting arsenic uptake primarily by lowering dissolved and labile arsenic in the soil rather than by accumulating more arsenic on root surfaces. The researchers attributed the contrast to differences in particle size, surface area, corrosion rate and iron-release kinetics among iron amendments. In particular, the relatively larger particles and lower specific surface area of mZVI may have limited direct contributions to rapid iron-plaque formation, while secondary iron minerals in the soil provided larger pools of arsenic-binding sites. Because the study did not measure the separate contributions of these pathways directly, that mechanism remains an interpretation supported by the combined observations.

Separating the treated soil into magnetic and non-magnetic fractions provided additional evidence that arsenic became associated with products of iron aging. The recovered magnetic material represented only 0.76% of the soil mass, yet it contained 127.7 milligrams of arsenic per kilogram, 3.23 times the concentration in non-magnetic soil, and 39.7 grams of iron per kilogram, 1.78 times as much. Magnetic particles retained 2.42% of the soil’s total arsenic and 1.35% of its total iron. Arsenic in this fraction was concentrated in the amorphous iron- and aluminum-oxide-associated fraction, which reached 96.1 milligrams per kilogram and was 11.2 times the corresponding concentration in untreated soil. The F4 concentration, representing arsenic associated with crystalline iron and aluminum oxides, reached 19.54 milligrams per kilogram in the magnetic fraction, compared with 14.7 milligrams per kilogram in the non-magnetic material. The results indicate that iron minerals formed as mZVI corrodes and ages can act as arsenic carriers, although the experiment did not establish how long that stability would persist under repeated field flooding and drainage.

The environmental cost appeared in the methane cycle. During anaerobic microcosm incubation, methane emissions began rising after 18 days with mZVI, compared with 24 days in controls, and peaked on day 50 at 108 micromoles per gram of dry soil. The control peaked at 86.4 micromoles per gram, making the amended peak 25% higher. At the end of anaerobic incubation on day 70, methane emissions were 31.4 micromoles per gram in controls and 36.6 micromoles per gram with mZVI. In the rice pots, porewater methane in the amended treatment reached 1,651 micromoles per liter at day 64 and subsequently remained within 1,522–1,651 micromoles per liter. Control porewater methane reached 2,890–3,166 micromoles per liter between days 79 and 94, whereas amended overlying water reached 191 micromoles per liter at day 79. The differing timing and concentrations mean the pot measurements cannot be read as a single uniform emission multiplier, but the microcosm experiment and microbial results together showed that mZVI shifted the balance toward net methane release. Aeration quickly reduced emissions in both treatments, and by day 78 the amended microcosms emitted 190 nanomoles per gram of dry soil per day versus 45.8 in controls.

Molecular and sequencing analyses identified changes in the organisms responsible for both methane production and methane consumption. In the 70-day microcosms, the abundance of the methanogenesis marker gene mcrA was approximately 2.49 times higher with mZVI, while the methane-oxidation marker pmoA was numerically higher but not significantly different from controls. In planted soil at maturity, both genes were significantly more abundant after amendment, reaching about 1.55 times the control level for mcrA and 1.43 times for pmoA. The amended soil also had higher bacterial richness and diversity, measured with the Chao1 and Shannon indices. Archaeal communities included Methanosarcina, Methanocella and Methanobacterium, while the type II methanotroph Methylocystis was enriched in treated soil. The balance suggests that methane oxidation did increase, but it was insufficient to offset stronger methanogenesis. More reducing conditions, relief from arsenic toxicity, altered organic-matter processing and potentially electron transfer through iron minerals were identified as possible contributors, but their individual contributions were not quantified.

The findings support mZVI as an effective laboratory amendment for arsenic immobilization and suggest that dry magnetic separation could recover some arsenic-bearing iron products, offering a partial route beyond stabilization alone. They also expose an unresolved sustainability problem: methane production was measured directly in soil incubation systems, not as complete field-scale emissions from flooded rice. Rice plants can transport methane through aerenchyma, and the study noted that plants may account for more than 90% of total emissions under some conditions, so the incubation measurements may not capture the dominant contribution of growing plants. The soil came from one field, each treatment had only three replicates, and arsenic speciation in soil and porewater, dissolved ferrous iron and total iron were not directly measured. These limitations prevented the researchers from quantitatively separating the roles of arsenic redox transformation, mZVI corrosion and iron reduction. Field testing with simultaneous measurements of arsenic species, iron transformations, plant-mediated methane transport and whole-system greenhouse-gas fluxes is needed before farmers can judge whether the contamination-control benefits outweigh the methane penalty.

Subject of Research: Environmental effects of microscale zero-valent iron in arsenic-contaminated paddy soil and rice systems

Article Title: Microscale zero-valent iron creates a trade-off between arsenic remediation and methane emissions in paddy soil–rice systems

Article References: Kong, A., Wang, X., Peng, X., Zhou, Y., Liu, J., Li, J., & Song, K. (2026). Microscale zero-valent iron creates a trade-off between arsenic remediation and methane emissions in paddy soil–rice systems. Frontiers in Environmental Science, 14, Article 1978292. https://doi.org/10.3389/fenvs.2026.1978292

Image Credits: AI Generated

DOI: 10.3389/fenvs.2026.1978292

Keywords: arsenic remediation, paddy soil, rice cultivation, methane emissions, zero-valent iron, methanogenesis, methanotrophs, soil biogeochemistry, iron minerals, greenhouse gas mitigation, Microscale, zero-valent