The research team, led by Kseniia Bondar of the Institute of Geophysics at the Polish Academy of Sciences, collected PBR as a magnetic extract from surface deposits at Stari Petrivtsi, a demolition site in the Kyiv region where controlled detonations of various munitions have been carried out since 2013. The magnetic fraction made up roughly one percent of the surface material by weight, so the researchers mixed the same proportion into natural Chernik Phaeozem soil sampled from an unploughed site in the Kharkiv region, replicating the contamination levels expected on frontline agricultural land subjected to prolonged shelling.

Under the scanning electron microscope, the residue revealed a two-part architecture. Angular metal fragments and spherical particles ranging from one to one hundred micrometres across were both built around a core of nearly pure metallic iron, sheathed in a corroded outer layer of iron oxides. Energy-dispersive X-ray analysis showed that the spherules carried a cocktail of potentially toxic elements, including copper, tin, lead, chromium, manganese and zinc, while the fragments additionally contained tungsten, molybdenum, aluminium, vanadium and antimony. Some spherules were dominated instead by antimony and lead, a reminder that munition alloys leave a varied chemical fingerprint in the ground.

To test how this debris behaves in a living soil, the team sealed the contaminated and control soils in containers with water, reproducing the waterlogged conditions that follow spring snowmelt or heavy summer rain on the heavy clay-rich chernozems of the region. For twenty days they tracked pH, redox potential, bacterial growth, dissolved organic carbon and the composition of gases in the sealed headspace, alongside colorimetric measurements of soluble ferrous and ferric iron. In the plain soil and the soil-plus-residue treatments, microbial metabolism stayed modest: oxygen drifted down from twenty-one to a few percent, pH eased from 7.8 to 7.0, and soluble iron never climbed above roughly fifty to eighty-five milligrams per litre.

The picture changed dramatically when the researchers added potato as a carbon and energy source, simulating the organic inputs that roots and crop residues provide in a real field. Fermenting microbes rapidly consumed the remaining oxygen, drove the redox potential from plus 420 down to minus 205 millivolts, and dropped the pH to 4.2. Hydrogen and carbon dioxide surged to fifty-eight and seventy-two percent of the gas phase respectively, while dissolved organic carbon jumped from 140 to 1200 milligrams per litre. Against this backdrop of intense anaerobic fermentation, soluble Fe(II) peaked at 960 milligrams per litre and Fe(III) at 420 milligrams per litre on day twelve, an order of magnitude above the low-activity treatments.

The mechanism the authors propose is a two-step partnership between biology and chemistry. Microorganisms cannot oxidise metallic iron directly, because the redox potential of the Fe(0) oxidation reaction sits below that of even hydrogen fermentation. Instead, iron metal is attacked abiotically by protons derived from water. What the microbes do is strip away the protective rust. Fermentation generates organic acids and protons that chelate and dissolve the iron oxide and hydroxide coatings on the fragments, exposing fresh metal and accelerating the chemical conversion of Fe(0) to soluble Fe(II). Vigorous gas release also stirs away reaction products from particle surfaces. In effect, the soil microbiome acts as a biocatalyst for the corrosion of its own contamination.

Magnetic measurements told a parallel story. Adding one percent PBR raised the soil’s mass-specific magnetic susceptibility by roughly thirty to fifty percent, and thermomagnetic analysis confirmed the presence of both magnetite and metallic iron in the residue. Yet waterlogging selectively dissolved the fine-grained, superparamagnetic magnetite produced naturally in the soil, halving the anhysteretic remanence in both control and contaminated samples. Unmixing of isothermal remanence acquisition curves showed that the highly coercive rust coating on the fragments, with coercivities near one hundred to one hundred sixty millitesla, decayed under waterlogged conditions, while the intermediate-coercivity signature of the spherules, between forty-seven and fifty-five millitesla, remained essentially untouched. After microbial stimulation, the coarse multidomain magnetite of the spherules came to dominate the entire magnetic spectrum, echoing the remarkable persistence of industrial fly-ash particles in peat bogs and lake sediments.

The biological consequences were assessed with seed germination and seedling growth assays using wheat and three Brassicaceae species: white mustard, radish and garden cress. Germination proved remarkably insensitive, staying above ninety-five percent in wheat, radish and cress regardless of contamination. Early seedling growth was a different matter. In radish, root elongation fell by eighteen percent and shoot growth by about twenty-one percent; in garden cress, roots shortened by twenty-two percent and shoots by nineteen percent, differences confirmed as statistically significant by Mann-Whitney tests. Wheat, by contrast, showed no measurable inhibition, and water extracts of the contaminated soil were nearly harmless, pointing to processes at the soil-root interface rather than bulk solution chemistry as the driver of toxicity.

Iron accumulation data reinforced the species-specific nature of the response. Wheat roots sequestered enormous quantities of iron, up to 6401 milligrams per kilogram in contaminated soil, while keeping shoot concentrations low, a classic excluder strategy that immobilises excess metal in the root apoplast and vacuoles. White mustard, in contrast, loaded its shoots with more than 1000 milligrams per kilogram of iron, whereas radish and cress actually transported less iron to their shoots under contamination, apparently throttling translocation through regulatory systems such as the IRT1 transporter and FRO2 reductase. Notably, the degree of growth inhibition did not track tissue iron levels, indicating that rhizosphere availability of Fe(II) during a transient window of mobilisation, roughly days eight to twenty-one of incubation, matters more than the iron ultimately stored in the plant.

The authors caution that their findings come from closed laboratory systems and must be validated in the field, where open mass exchange and repeated wetting cycles could amplify or dilute these effects. Even so, the study offers the clearest picture yet of how war debris ages in fertile farmland: rusted fragments slowly dissolve under microbial pressure, releasing pulses of bioavailable iron that stress sensitive seedlings, while the explosion-derived spherules persist as a durable magnetic fingerprint. For the vast agricultural regions of Ukraine awaiting demining and recovery, the message is that the environmental legacy of shelling is not static. It is a living geochemical process, driven by the smallest inhabitants of the soil, and its full consequences for crop safety and food-chain transfer are only beginning to come into focus.

Subject of Research: Microbial and magnetic transformation of post-blast weapons residue and its phytotoxic effects in Ukrainian chernozem soils

Article Title: Fate of post-blast residue in soils from an armed conflict zone determined by magnetic, microbiological, and ecotoxicological assessments

Article References: Fate of post-blast residue in soils from an armed conflict zone determined by magnetic, microbiological, and ecotoxicological assessments. (n.d.). https://doi.org/10.1007/s10653-026-03497-x

Image Credits: AI Generated

DOI: 10.1007/s10653-026-03497-x

Keywords: post-blast residue, soil contamination, armed conflict, iron mobilization, soil microbiome, magnetic susceptibility, phytotoxicity, chernozem, waterlogging, seedling growth, Ukraine, biogeochemistry