Odontarrhena chalcidica is the workhorse of European nickel agromining: an obligate serpentine endemic whose foliage can run to around 2% Ni by dry mass in the field. Whether its extraordinary uptake is mined — root exudates and rhizosphere microbes actively dissolving Ni-bearing minerals — or merely harvested from the Ni already sitting in the labile soil pool is a live argument in the literature, and it matters for how agromining soils should be managed. Trimmel, Irrgeher and colleagues at Montanuniversität Leoben and BOKU (Vienna) have now answered it with an instrument bulk chemistry cannot argue with: they synthesised a Ni-bearing saponite clay in which a known fraction of the Ni was the rare stable isotope ⁶¹Ni, mixed it into two real ultramafic soils, and watched where the label went (Anal Bioanal Chem 418:4481–4495, 2026, CC BY).
The uptake that bulk chemistry missed
The setup was a 14-day RHIZOtest: roughly 2 g of soil per unit, amended with 0.3 g of either natural ground serpentinite, unlabelled saponite, or ⁶¹Ni saponite, with five replicates per group. The two soils came from an ultramafic forest at Redlschlag, Burgenland, Austria, along a natural gradient: soil S1 holds 552 ± 52 mg kg⁻¹ aqua-regia Ni (41.6 ± 0.5 DTPA-extractable), while soil S6 holds 1,465 ± 58 (158 ± 7). On paper, nothing happened. Total Ni in planted versus unplanted soils showed no significant mobilisation; DGT-labile Ni — the kinetically available pool that approximates what a root can pull — did not differ between planted and unplanted units in any treatment; and plant Ni concentration did not correlate with c_DGT(Ni) at all. Yet isotope pattern deconvolution of the shoot digests told a different story: on the low-Ni soil S1, 19.3 ± 5.0% of the Ni in the shoots came from the amendment, versus 7.7 ± 1.8% on S6 (Welch’s t-test, p = 0.00486). In absolute terms, plants on S1 carried 6,700 ± 3,400 ng g⁻¹ of amendment-derived Ni in their shoots against 2,400 ± 1,100 on S6 — nearly three times more, from the same amendment. The authors’ own verdict: the difference “would not have been detectable through total quantitative Ni measurements alone.”
The amendment over-delivers to the labile pool
The DGT eluates explain why the plant could find the label even though bulk readings stayed flat. The amendment contributed only ~0.66% of pseudo-total Ni in the S1 mixture (0.27% in S6), but 1.5 ± 0.1% of the DGT-labile Ni in planted S1 (1.6 ± 0.1% unplanted; S6: 0.54 ± 0.05 planted, 0.47 ± 0.05 unplanted). Amendment Ni is thus roughly two-and-a-half times overrepresented in the pool a root actually draws on — though still below the 3.7% (S1) and 1.8% (S6) you would predict from DTPA extraction, meaning the freshly synthesised clay feeds the chelator-accessible pool even faster than the DGT-accessible one. Notably, the synthetic saponite itself was agronomically bland: shoot biomass did not differ across treatments (p = 0.296 for S1, p = 0.207 for S6), so the uptake signal is not a stress artefact.
A detector sensitive to 5 picograms per gram
The reason this works is sensitivity. In standards of 50 ng g⁻¹ total Ni, the group could still resolve a ⁶¹Ni spike at a molar fraction of 0.01% — about 5 pg g⁻¹ of ⁶¹Ni, a ~5‰ shift in the Ni isotope ratio, nearly 200 times the analytical uncertainty — and at 1.5% spike the shift reaches 500–600‰. The spiked saponite itself carried a measured ⁶¹Ni molar fraction of 4.54 ± 0.01% in total digests (4.71 ± 0.04% in EDTA extracts) against a theoretical 5.92%. That dynamic range lets a two-week experiment see source partitioning that months of conventional monitoring miss: earlier work on these same soils found no rise in porewater Ni until day 49. The authors are careful not to overclaim — a 14-day window may simply be too short for rhizosphere processes to dent the bulk labile pool, and root-tip-localised mobilisation documented by earlier imaging work would sit below DGT’s spatial footprint either way.
What it means for phytomining — and the honest bounds
The non-obvious implication runs in both directions. Opportunity: on marginal, Ni-lean ultramafic land — the kind that dominates the land bank available to agromining — a plant appears to lean harder on an added Ni source, drawing ~2.5× more of its shoot Ni from the amendment than plants on rich soil, which suggests Ni-bearing amendments (tailings, lateritic residues, reactive mineral phases) are worth real agronomic testing on poor sites, and this tracer is the tool that can prove the metal actually entered the crop rather than redistributing in the soil. The same method transfers directly to tracing the environmental fate of metal-bearing waste materials. Risk: the absolute effect here is small and the bounds are real — 14 days, ~15% amendment by mass, n = 5, shoot Ni of 33,000–35,000 ng g⁻¹ (i.e. 33–35 µg g⁻¹, far below the 1,000 µg g⁻¹ hyperaccumulation threshold these young, pre-rooted plants would eventually reach), variances nearly as large as the means, and roots never analysed. And whether the low-Ni plant released more amendment Ni or merely took up more of what was released “remains to be investigated” — the paper’s words, and the right caveat.
Source: Trimmel, Epov, Abu Zahra, Berger, Prohaska, Puschenreiter, Siebenbrunner, Tognacchini, Wagner & Irrgeher (2026), Analytical and Bioanalytical Chemistry 418(14):4481–4495, DOI 10.1007/s00216-026-06539-6 (CC BY 4.0, published 2026-05-02). Every load-bearing number was verified verbatim against the full text by three independent reviewers via two independent retrieval paths (Europe PMC full-text XML and the PubMed Central HTML page); all values agreed. Noted source-level caveats: the detection-limit paragraph contains an isotope-ratio index typo (⁶¹Ni/⁶⁰Ni vs ⁶¹Ni/⁶²Ni — the 5.136 ± 0.028‰ value itself is unambiguous), and one parenthetical in the Conclusion contradicts the Results on planted-vs-unplanted DGT differences; we follow the Results. See /methodology/ for how we source and check analysis pieces.