Chemically assisted phytoextraction keeps looking for an alternative to EDTA — something that mobilises metals without leaving a persistent, leachable chelant in the soil. Petrov, Stefanova, Daskalova, Todorova, Miteva and Mihailova (University of Forestry, Sofia) have tested one of the more elegant candidates on a genuinely difficult material: halide-assisted mobilisation, the same ligand chemistry behind induced gold accumulation, applied not to gold but to 27 elements at once in mustard (Brassica juncea) and common chicory (Cichorium intybus) grown on a metal-rich post-mining substrate (Processes 14(18):2910, 2026, gold open access). The headline is not that it worked. It is how it worked: element by element, species by species, and with the strongest response at the lowest dose.
A screening design on real technogenic material
The substrate came from a historical waste-rock dump in the Srednogorie mining region of Bulgaria, sampled at 750 m elevation in February 2026 and bulked from five points at 30–50 cm depth. It is seriously contaminated: Pb ranged 739.7–1022 mg kg⁻¹, Ba 649.1–720.1, Zr 194.0–273.1, Cu 173.0–255.5, Cr 161.4–184.4, Rb 154.5–171.1, Zn 102.5–171.8 and V 62.7–83.5 mg kg⁻¹ dry weight — and strongly acidic (pH 2.6), so every pot was first limed with 2.47 g Ca(OH)₂, which lifted pH to 7.3 alongside NPK fertilisation.
Into this went 60 pots — 30 per species, 300 g substrate each — with potassium iodide (KI), calcium chloride or ammonium chloride applied once at 20, 50 or 80 mg kg⁻¹. Two design choices matter for reading everything that follows. First, citric acid (1 g kg⁻¹) went into all pots, controls included, as a root-exudate-like background — so every treatment effect is the salt’s effect on top of citric acid, and the study has no citric-acid-free baseline. Second, the three pots per treatment were pooled into a single composite sample, and concentrations were measured by semi-quantitative portable XRF. There are no replicates, no inferential statistics, and no biomass measurements: everything below is a concentration ratio, not a removal rate.
The mechanism: halide chemistry, borrowed from gold phytomining
The reagents were chosen to probe halide-assisted metal complexation — iodide and chloride form soluble complexes with several metals, the principle behind the classic induced accumulation of gold, and chloride leaching is standard hydrometallurgy. The clearest response was vanadium, and it was striking. In chicory, root bioconcentration of V (BCF_root, leaf-root tissue over substrate) went from 0.099 in the control to 5.68 under KI at 20 mg kg⁻¹ — a 57.2-fold increase — then fell to 2.18 and 2.41 at the higher doses. The strongest response was at the lowest dose.
In mustard the same pattern held in roots (BCF_root 4.08, 3.37, 2.23 under KI 20/50/80 against a censored control below 0.016), and at KI 80 mg kg⁻¹ something more interesting happened: V shoot bioconcentration reached 3.60 and the root-to-shoot translocation factor reached 1.62 — the one treatment in the whole experiment that produced both tissue enrichment relative to the substrate and genuine transfer to harvestable leaves.
The authors’ mechanistic hypothesis, stated as a hypothesis, is chemically plausible: under the near-neutral, oxic conditions of the conditioned substrate, vanadium should sit mainly as the vanadate oxyanion H₂VO₄⁻/HVO₄²⁻ — a structural phosphate analogue that enters roots through phosphate pathways. Iodide is redox-active and could partially reduce V(V) to the more soluble vanadyl cation V(IV), while the accompanying K⁺ may perturb phosphate–vanadate competition at the root surface. Speciation was not measured, so this remains a mechanism to be tested, not a result.
Species identity beat reagent chemistry
The strongest determinant of accumulation was not the reagent but the plant. Mustard more often retained Y, Sr, Th and Zn in its roots — consistent with Brassica’s phytochelatin-and-vacuole sequestration machinery — while chicory translocated Rb, Sr, Pb and Cu to its leaves more readily. For a remediation practitioner this is the useful decomposition: root retention is phytostabilisation (nothing leaves the site unless you excavate roots), leaf transfer is phytoextraction (harvestable removal). A single species will not do both jobs.
The yttrium result is the cautionary one for agromining. Y — a critical raw material — was enriched in roots (BCF_root up to 4.97 in chicory and 5.67 in mustard, above 1 in nine of ten treatments in each species) but its translocation factor generally stayed below 1. Non-specialist plants take up REEs and then lock them below ground, exactly as the wider REE-phytomining literature has found; concentration in roots is not a recovery route. Anyone reading pot-scale BCF values as a phytomining prospectus should sit with that contrast — see our rare-earth-elements hub for the state of that field.
The translocation-factor trap
The paper’s most transferable lesson is methodological. Of the 385 treatment–element–index combinations with exact control-relative values, 170 increased and 215 decreased — no reagent was a general enhancer. Of 146 complete three-dose series, 88 were non-monotonic and only 25 increased monotonically; dose escalation bought nothing, systematically.
Worse, big translocation-factor gains could be root losses in disguise. In chicory under KI 50, Y’s TF rose 10.1-fold — but its BCF_root had collapsed to 0.29 of control, so the leaves were relatively richer mostly because the roots were poorer. In mustard under NH₄Cl 50, Y’s TF rose 8.40-fold while BCF_root fell to 0.18 of control and shoot BCF stayed below 0.4. And the single largest TF in the dataset — 208.96 for Zn in the mustard control — was an artefact of a 1.15 mg kg⁻¹ root concentration meeting a 240.3 mg kg⁻¹ leaf concentration, which the authors themselves quarantine as unstable. A TF above 1 tells you about allocation between tissues, not about enrichment relative to soil; read together with both BCFs or not at all.
What it means — and what it does not
The opportunity: a cheap, non-persistent salt class that, for specific element–species pairs, produces large enrichment responses on real degraded material — V in mustard under KI being the standout — at doses (20 mg kg⁻¹) far below chelator rates. The genuine risks are exactly the ones the authors flag: mobilised metals that outlive uptake window become a leaching liability, so the lowest effective dose is the defensible one; the citric-acid background means the absolute mobilisation is unknown; and with no biomass data, no replication, semi-quantitative XRF and one potting round, nothing here quantifies removal per hectare or recovery economics. This is a screening-stage study, honestly labelled as such by its authors, and the right next step is their own: replicated validation with untreated and citric-acid-only controls, pore-water and leachate chemistry, biomass and yield, and repeated harvest cycles. Vanadium has no hub on this site — this piece is the first time we have covered it — precisely because field-grade numbers for it are this rare.
Provenance: all figures from Petrov et al. (2026), Processes 14(18):2910, doi:10.3390/pr14182910 (gold OA, full text read); every load-bearing number independently verified against the primary source by three reviewers on three model vendors with zero discrepancies. See /methodology/ for how we source and check analysis pieces.