Taylor 2016 — Arsenic concentration and speciation across 20 New England seaweed species

Taylor and Jackson surveyed total arsenic and arsenic speciation across 20 macroalgae species (46 samples; plus flowering plants, mixed products, and fertilizers) harvested for food and agriculture in New England, by ICP-MS and HPLC-ICP-MS. Total arsenic ranged 4.1–106 mg/kg, occurring mostly as arsenosugars, and inorganic arsenic was a minor component (<1 mg/kg) across nearly all taxa — with one striking exception: the kelp Laminaria digitata from Maine carried inorganic arsenic of 2.8–20.4 mg/kg (mean 8.32 ± 7.67), and two mixed granulated seaweed products containing this species held 6.6 and 10.9 mg/kg. Within this survey it is L. digitata kelp, not hijiki, that accumulates toxicologically significant inorganic arsenic, and the elevation is harvester-/region-variable even within one species (the single New Hampshire L. digitata sample was <0.01 mg/kg iAs).

Key numbers

Total arsenic (As T) and inorganic arsenic (iAs) in mg/kg dry weight (air-dried); from Table 1. <MDL = below method detection limit. Means ± SD where n>1.

Brown algae (Ochrophyta):

Red algae (Rhodophyta), total As 4.17–20.73 mg/kg; iAs all <0.5 mg/kg: Porphyra umbilicalis (ME) 20.73 / iAs 0.12; Heterosiphonia japonica (NH) 8.24 / 0.47; Polysiphonia lanosa (NH) 14.00 / 0.26; Chondrus crispus (NH, n=2) 12.13 ± 0.97 / 0.07 ± 0.08; Chondrus crispus (ME) 6.10 / 0.06; Phyllophora pseudoceranoides (NH) 4.17 / 0.14; Gracilaria vermiculophylla (NH) 11.80 / 0.23; Palmaria palmata (ME, n=4) 8.95 ± 4.80 / 0.06 ± 0.11.

Green algae (Chlorophyta), total As 4.13–14.65 mg/kg; iAs all <0.2 mg/kg: Chaetomorpha picquotiana 6.69 / 0.09; Gayralia oxysperma 12.68 / 0.17; Ulva lactuca (NH) 5.34 / 0.13; Ulva lactuca (ME) 4.13 / 0.02; Ulva prolifera 14.65 / 0.12.

Mixed and other:

• Mixed granulated seaweed product “soup1” (ME): total As 57.29, iAs 6.56; “soup2” (ME): total As 49.03, iAs 10.92 — both contain L. digitata.
• Aquatic flowering plants (lower than macroalgae): Zostera marina (NH, n=2) total As 4.42 ± 1.12, iAs 0.27 ± 0.08; Spartina alterniflora (NH) 3.40 / 0.10.
• Liquid seaweed fertilizers (A. nodosum-based): total As 2.29 and 5.76 mg/kg; iAs 0.08 and 0.58.

Additional source-reported facts:

• Total As 4.1–106 mg/kg overall; arsenosugars made up the dominant water-extractable species (>85% of soluble As in most seaweeds); methylated species (mostly DMA, trace MMA) and trace arsenobetaine present in several samples.
• Total As significantly higher in brown algae (Ochrophyta) than red/green; the order Laminariales had the highest As and the widest range (13.9–106 mg/kg). Brown-algae As correlated with phosphorus (r²=0.45, p<0.001).
• A partial-speciation method (1% HNO3 microwave extraction) gave inorganic-arsenic results comparable to the methanol:water method (iAs cross-method regression y=1.07x+0.61, r²=0.89), supporting the simpler approach for monitoring.
• Exposure: a 5 g portion of the high-iAs L. digitata (2.8–20.4 mg/kg iAs) gives a dose of 0.19–1.4 µg iAs/kg body weight for a 75 kg adult — half to more than four times the US EPA oral reference dose (0.3 µg iAs/kg bw/day) — and exceeds drinking 1 L of water at the 10 µg/L limit.
• Regulatory context cited: US EPA oral RfD 0.3 µg iAs/kg bw/day; EU iAs limits 0.2 mg/kg (white rice) and 0.1 mg/kg (rice-based infant/child foods); EU animal-feed limit 40 mg/kg total As (assuming iAs <2 mg/kg); US sewage-sludge land-application limit 41 mg/kg total As. Total As in Laminaria exceeded the sewage-sludge and fish-feed levels.

Methods (brief)

Seaweed was obtained from 7 commercial harvesters in northern Maine and hand-collected at two Great Bay Estuary sites in New Hampshire (December 2012), air-dried, and homogenised. Total As (and S, Fe, P) was determined after concentrated-HNO3 high-pressure microwave digestion (MARS XPRESS) by collision-cell ICP-MS (Agilent 7700x, He collision gas); S/Fe/P by magnetic-sector ICP-MS (Element 2). Arsenic speciation used a 1:1 methanol:water extraction (sonication + overnight) for all samples, with a subset of 30 also extracted by a partial-speciation method (1% HNO3 + microwave heating 55/75/95 °C, then H2O2) to capture inorganic As; separation by anion-exchange HPLC (Agilent LC1120, Hamilton PRP-X100, 20 mM ammonium carbonate pH 9), with cation-exchange (Supelcosil SCX) to verify cationic species. Quality control used CRMs DORM-2 (dogfish; AB/TMAO), an in-house kelp standard with a consensus iAs value (Brooks Rand interlaboratory comparison), Antarctic-krill MURST-ISS-A2, and NIST 1568a rice flour (iAs); total-As recoveries 95 ± 3% (DORM-2), 86 ± 2% (MURST-ISS-A2), 111 ± 6% (NIST 1568a). The authors note that round-robin iAs determination in algae performs poorly (most labs overestimate iAs from peak overlap), and that mild extractants give low/variable efficiency while acidic extraction collapses arsenosugar side chains; the partial-speciation method still yields accurate total iAs. All values dry weight.

Implications

• Certification (HMTc): a direct-evidence occurrence-and-speciation source for the Category 6 seaweed-kelp-foods row (iAs/tAs platform), showing that the kelp Laminaria digitata — a core component of the row — can carry toxicologically significant inorganic arsenic (2.8–20.4 mg/kg DW, plus 6.6–10.9 mg/kg in L. digitata-containing mixed products) while every other surveyed taxon (Ascophyllum, Fucus, Saccharina, Alaria, red and green algae) stays <1 mg/kg iAs. By the paper’s own exposure calculation, a 5 g portion of the high-iAs L. digitata delivers half to more than four times the US EPA inorganic-arsenic oral reference dose. This reinforces a kelp-specific (especially Laminaria) high-iAs sub-treatment within the row’s clean/dirty structuring, and shows the elevation is harvester-/region-variable even within one species (Maine high; the single NH L. digitata <0.01 mg/kg). All on the dry-weight, as-harvested basis.
• Courses: a teachable methods case on why speciation (not total As) is required for seaweed monitoring and why partial-speciation methods are fit for purpose.
• App: contributes to the seaweed ingredient contamination_profile for total and inorganic arsenic, with the Laminaria-kelp high-iAs caveat.

Verification notes

• raw_handle MFK_taylor2016 from the PDF filename; raw_path under “raw/Manual Fetch Kimi /June 8 Inorganic Arsenic Seaweed/“. DOI 10.1016/j.chemosphere.2016.08.004 confirmed on the article header. This is the same study cited by Ronan 2017 as the New England comparator.
• Evidence tier A: peer-reviewed, ICP-MS / HPLC-ICP-MS speciation with multiple CRMs and a partial-speciation cross-validation.
• Speciation: total As and inorganic As reported and kept distinct (lifted to frontmatter as tAs, iAs). Organic species (DMA, MMA, arsenobetaine, four arsenosugar analogs) were measured and are summarised in the body but NOT lifted to frontmatter — they are not HMI analyte-vocabulary tokens and are not HMTc-relevant analytes; recording them as iAs would be a speciation error, which is avoided.
• Units/basis preserved exactly (mg/kg dry weight, air-dried). <MDL recorded as such (censored), not as a numeric value. The key L. digitata (ME) iAs is given both as the Table 1 mean ± SD (8.32 ± 7.67) and the Results-text range (2.8–20.4 mg/kg).
• sample_n=46 is the macroalgae sample count summed from the Table 1 n-column (brown 29 + red 12 + green 5); these span 20 distinct macroalgae species across 25 species-site rows. Non-macroalgae samples (3 flowering plants, 2 mixed products, 2 fertilizers) are described in sample_population, not counted in sample_n. The abstract’s “10 species” refers to the Great Bay collection subset; the full survey spans more species via the commercial harvesters.
• Brand firewall: not engaged. Mixed products are labelled generically (“soup1”, “soup2”); fertilizers as “A. nodosum fert” / “fert fish”; commercial harvesters are not named. No commercial brand names or brand-by-brand contamination rankings.
• Jurisdiction US (Maine + New Hampshire); sampling_locations records the two regions; sampling_year_range 2012.
• matrices [edible-seaweed, macroalgae, kelp, dry-weight] per corpus convention; kelp directly warranted (Laminaria, Saccharina, Alaria, Agarum — order Laminariales).
• Instrument/CRM/vendor names (Agilent, CEM MARS, Thermo Element 2, Hamilton, Supelcosil, DORM-2, MURST-ISS-A2, NIST 1568a) retained in Methods as permitted scientific reporting.
• Audit subagent (2026-06-08, fresh-context) returned REVISE; all Table 1 values and the iAs/tAs speciation labeling (highest-stakes) verified exact (the L. digitata ME iAs 8.32±7.67 correctly taken from the iAs column, not the AsSugPO4 10.63±8.71 column). Three findings applied: (1) the taxa count “24” was wiki-introduced and unsupported — corrected to 20 distinct macroalgae species / 25 species-site rows / 46 samples throughout; (2) the “ANZ 1 mg/kg / French 3 mg/kg seaweed limits” comparison was removed from Implications because Taylor 2016 does not state those limits (they are cited in ronan2017 and fsanz2012 instead) — replaced with the paper’s own EPA-RfD exposure finding; (3) the cross-source “reproduces / second independent confirmation (Ronan 2017)” synthesis claim was removed from the body and Implications (cross-source synthesis is the Part 9 workflow, not the source page). The provenance fact that Ronan 2017 cites this study as its New England comparator is retained below as a neutral note.
• Provenance: Ronan et al. 2017 (Chemosphere) cites this study as the New England comparator for L. digitata arsenic; the cross-study comparison itself belongs to the synthesis workflow, not this page.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.