Fertiliser is the cheapest thing a phytoextraction operator can add, so “feed the crop to mine more metal” sounds like free money. A new open-access pot study prices the trade. On a strongly acidic, cadmium-contaminated soil (total Cd 8.40 mg kg⁻¹, pH ~4.6), Li, Li, Liu, Deng and Xiao (2026) in Agronomy grew grain amaranth (Amaranthus hypochondriacus cv. ‘K112’) under seven phosphorus regimes — soil-applied KH₂PO₄, CaHPO₄ and NH₄H₂PO₄, plus foliar KH₂PO₄ at 0.2/0.4/0.8% — four replicates each, 28 pots in a greenhouse. Soil P worked: total plant cadmium accumulation rose +69.8% (KH₂PO₄), +105.6% (CaHPO₄) and +139.5% (NH₄H₂PO₄) over the unfertilised control. The interesting part is how it worked, because the mechanism differs by fertiliser and one of the winners carries a hidden bill.
The gain is mostly biomass, not freed metal
The obvious hypothesis — phosphate mobilises soil cadmium, plant takes up more — is the one the data reject. Bulk plant-available cadmium (0.01 M CaCl₂-extractable) did not differ significantly across any treatment (control 2.86, NH₄H₂PO₄ 2.82 mg kg⁻¹). Under soil KH₂PO₄ the exchangeable pool actually fell — BCR-extraction F1 down 7.8%, with cadmium shifting into more recalcitrant fractions — and per-tissue root concentration dropped, yet total removal still rose +69.8%. In that case the fertiliser bought biomass, not concentration: more kilograms of shoot, each carrying about the same cadmium, integrate to a larger harvest.
This is consistent with — though not the same mechanism as — the labile-pool ceiling we reported for sunflower on 30 July. There, uptake collapsed across cycles as the available fraction was drawn down over time; here the available pool is statically flat within a single cycle and biomass does the lifting. Both land on the same operator’s warning — the total (or bulk-available) metal on the soil report is not what governs the harvest — but this is a one-cycle snapshot. Nothing in it tests whether amaranth could sustain +139% into a second cropping, and it should not be read that way.
The strongest lever is acid, and the assay can’t see it
NH₄H₂PO₄ is the exception that complicates the tidy “it’s just biomass” story. It gave the highest cadmium in every tissue, pushed the shoot translocation factor from 1.26 to 2.10 and the shoot bioconcentration factor from 10.4 to 13.7, and lifted leaf cadmium ~38%. It did this by acidifying the soil — pH down 0.24 units (4.54 → 4.30) from ammonium nitrification, available phosphorus up 175%. So here uptake efficiency and root-to-shoot translocation genuinely rose — while the CaCl₂-available pool stayed flat. Read carefully, that is not evidence that availability is irrelevant; it is evidence that a single-shot CaCl₂ assay failed to capture the rhizosphere/kinetic mobilisation that acidification drove. The two “labile” proxies in the paper even disagree with each other (CaCl₂ unchanged; BCR-F1 moved), which is a sharper caution than either alone: no single soil-availability number in this study predicts total phytoextraction.
The bill: an acidified soil handed to the next crop
The authors are candid that the best lever is the least sustainable one: repeated NH₄H₂PO₄ “may intensify soil acidification and increase heavy-metal mobility.” But the leaching risk is genuinely two-sided, not a clean consequence — the same phosphate additions pushed a large share of soil cadmium into the residual fraction (F4 up ~75–150%, i.e. immobilisation) even as protons free the exchangeable margin. This closed pot measured no leachate, so which way the balance tips in the field is untested; treat off-site mobilisation as a flagged, unquantified hazard, not a result.
The real, specific hazard is subtler and is our inference, not the paper’s: because the loaded biomass is harvested away, the exposure is not the amaranth crop but the legacy soil — more acidic and more cadmium-labile — left for whatever food crop follows in rotation. And because cadmium here concentrated in leaves (control leaf 105.8 ≫ root 61.0 ≫ stem 52.1 mg kg⁻¹) and grain cadmium was never measured, this study says nothing about the edibility of amaranth grain grown this way — only about the state of the ground afterward.
Is it really a cadmium hyperaccumulator? Not in the field sense
The paper states that leaf cadmium exceeded 100 mg kg⁻¹ in all treatments and that shoot TF and BCF both exceeded 1, “indicating that this species met the criteria for a Cd hyperaccumulator.” We dissent, on definition rather than arithmetic. The 100 mg kg⁻¹ shoot threshold (van der Ent et al. 2013; Reeves et al. 2018) implies extraordinary uptake relative to a normal substrate — not a value reached only on soil spiked to 8.40 mg kg⁻¹, where a shoot BCF of ~10 mostly reflects a small available-cadmium denominator. Grain amaranth does not appear in the global hyperaccumulator database, and the paper’s own discussion elsewhere calls it a “high-biomass accumulator.” That is the honest label: a fast, high-yield facultative accumulator and phytoextraction crop, not a constitutive metallophyte in the sense of Noccaea caerulescens or Sedum alfredii, and unlike the true Amaranthaceae cadmium accumulator Gomphrena claussenii in our database, its value is biomass × moderate uptake, not exceptional physiology.
The lever the design can’t see
Because fertiliser form and soil pH were never varied independently — the authors flag the absence of pH-buffered controls — NH₄H₂PO₄’s “win” may be little more than you added acid. If the phosphate itself tends to stabilise cadmium (the F4 rise) while the acidification does the mobilising, then the actionable move is to split the two: take the biomass and P-nutrition benefit from a non-acidifying source — CaHPO₄ raised pH and still delivered +105.6% — and manage soil pH deliberately and reversibly (e.g. controlled acidulation during an extraction window, liming on exit) rather than baking irreversible acidification into a repeated ammonium-phosphate regime on land destined to grow food again.
Bottom line. This is one greenhouse pot trial, one soil, one cultivar (n = 4) — bound the claims accordingly. Its defensible contribution is not a new metal crop but a clean dissection: phosphorus lifts cadmium phytoextraction by grain amaranth mainly by growing more plant; the one lever that also raises uptake does so by acidification the bulk assay cannot see; and that lever trades a bigger harvest for a more labile, more acidic soil whose leaching balance and food-crop legacy this closed system never measured.
Provenance: every figure here was read from the open-access full text (CC BY) and independently re-verified against the source by three separate reviewers; the definitional dissent and the flagged risks are ours. See our methodology for how we source and check.