Hovanec 2004 — Arsenic speciation in peanut butter by IC-ICP-MS

This 2004 paper from West Coast Analytical Service (Santa Fe Springs, CA) is the first published arsenic speciation study on commercially available peanut butter. Six retail peanut butter spread samples (creamy, super-chunk, low-fat spread, single-serve cups, crunchy salted), plus one whole-peanuts-in-shell sample and NIST SRM 2387 peanut butter reference material, were measured for total arsenic by ICP-MS. Three of the peanut butter spreads plus SRM 2387 were subjected to full four-species IC-ICP-MS speciation (As(III), DMA, MMA, As(V)). The paper developed an extraction protocol using acetone pre-extraction followed by 50% 2-butoxyethanol, solving the problem of incomplete arsenic recovery from high-fat matrices that had caused prior methanol-water methods to fail. Critically, the dominant arsenic species across the speciated samples were organic (DMA and MMA), with inorganic species detectable in only one of the three peanut butter spreads tested, suggesting lower toxicological concern than matrices dominated by iAs.

Key numbers

Total arsenic (ng/g, wet weight; Table 3):

• JCPB1 (creamy): 19.0 ng/g
• GCPB1 (creamy): 7.5 ng/g
• SCPB1 (super chunk): 21.4 ng/g
• BPB1 (low fat spread): 6.5 ng/g
• SPPB1 (creamy 1 oz cups): 6.9 ng/g
• TJPB1 (crunchy salted): 14.4 ng/g
• WPPB1 (whole peanuts in shell): 9.9 ng/g
• SRM 2387 (NIST peanut butter): 12.0 ng/g
• Detection limit: 0.4 ng/g
• Total-As range across the six peanut butter spreads: 6.5–21.4 ng/g

Speciation of JCPB1 (creamy, total As = 19.0 ng/g; Table 6):

• As(III): 1.6 ng/g
• DMA: 6.1 ng/g
• MMA: 4.9 ng/g
• As(V): 6.3 ng/g
• Sum of species: 18.9 ng/g (agrees with digestion total, −0.5% difference)
• Duplicate sum: 18.2 ng/g (−4.2% difference)
• Inorganic As (As(III) + As(V)): 7.9 ng/g (~41% of total As)

Speciation of SCPB1 (super chunk, total As = 21.4 ng/g; Table 6):

• As(III): ND
• DMA: 14.8 ng/g (16.1 ng/g duplicate)
• MMA: 7.9 ng/g (9.1 ng/g duplicate)
• As(V): ND
• Speciated total: 22.7 ng/g (+6.1%); duplicate 25.2 ng/g (+17.8%)
• Inorganic As: not detected

SRM 2387 (NIST peanut butter, total As = 12.0 ng/g; Table 6):

• As(III): ND
• DMA: 8.9 ng/g (first replicate); 7.6 ng/g (duplicate); abstract reports 8.3 ng/g average
• MMA: 5.5 ng/g (first replicate); 3.9 ng/g (duplicate); abstract reports 4.7 ng/g average
• As(V): ND
• Speciated total: 14.4 ng/g (+20%); duplicate 11.5 ng/g (−4.2%)
• Inorganic As: not detected
• These are the first published total-As and speciation values for NIST SRM 2387 (no certified As value at time of publication)

Spike recoveries (Table 5):

• JCPB1 (50 ng/g spike): As(III) 93.2%, DMA 89.6%, MMA 87.4%, As(V) 98.4%
• TJPB1 (18 ng/g spike, near detection limit): As(III) 122%, DMA 76.0%, MMA 54.0%, As(V) 40.0%

Methods (brief)

Analytical method: ICP-MS for total As (PerkinElmer ELAN 6100, no collision cell, AS91 autosampler, FIAS 400) after heated HNO3/H2O2 digestion at ~120 °C for 90 min on 1 g ± 0.2 g aliquots, with 3% methanol added to all standards and digests for signal enhancement. IC-ICP-MS (ELAN 6000) for speciation after acetone de-fatting (20 mL, vortexed, centrifuged 6000 rpm × 20 min, supernatant discarded) followed by two cycles of 50% 2-butoxyethanol–water extraction (10 mL each, 60 °C reciprocating water bath × 30 min, centrifuged, supernatants combined). LOD for total As: 0.4 ng/g (includes 50× dilution factor). Four arsenic species (As(III), DMA, MMA, As(V)) separated on a Hamilton PRP X-100 anion-exchange column (250 × 4.1 mm, 10 µm) with 30 mM NH4H2PO4 + 3% MeOH mobile phase at pH 6.5, 1.2 mL/min isocratic, 13 min run. Enriched Se-78 internal standard added post-extraction to monitor column efficiency and correct retention-time and signal variation. IC-HG-ICP-MS used as an orthogonal verification mode (without UV photo-oxidation, allowing differentiation of As(III) from arsenobetaine and MMA from arseno-sugars) and confirmed no co-eluting peaks or interfering analytes.

Key limitations: small n = 6 commercial peanut butter spread samples (plus one whole-peanut and one SRM), single laboratory, 2004 vintage. Speciation was performed on only three of the six spreads (JCPB1, SCPB1, TJPB1) plus SRM 2387. Inorganic-As species (As(III) + As(V)) were detected only in JCPB1; SCPB1 and SRM 2387 showed only the methylated organics. Speciated totals diverge from digestion totals by up to ~20% (SRM 2387), reflecting matrix and recovery effects. Spike recoveries for TJPB1 were lower (40–122%) due to the spike concentration sitting near the solution detection limit. Product formulation differences (oil separation, salt content, organic vs conventional sourcing, crunchy vs creamy) may affect speciation; mechanism not fully explained.

Implications

Certification: Total As in peanut butter ranges 6.5–21.4 ng/g (ppb) on a wet-weight basis. Inorganic arsenic appears to be a minority species, and was undetectable in two of the three peanut butter spreads tested in full speciation (SCPB1, plus NIST SRM 2387). However, the iAs fraction varies substantially between products (0% to ~41% of total As in this dataset), so total-As measurement alone is insufficient for speciation-based risk assessment. HMT&C peanut-butter or nut-butter product pages should note both the absolute concentration range and the variable iAs fraction. Sample size is small (n = 6 spreads, n = 3 speciated) and dates to 2004; pair with more recent peanut and tree-nut speciation studies before setting thresholds.

Courses: Demonstrates why arsenic speciation matters for risk characterisation: a product with 20 ng/g total As but predominantly DMA/MMA has very different toxicological implications than the same total As dominated by iAs. Also demonstrates a methodological lesson — high-fat food matrices (peanut butter, edible oils) are not amenable to standard methanol-water arsenic extraction; 2-butoxyethanol with acetone pre-extraction was required to recover the spike. Good teaching case for both the speciation module and the extraction-method module.

App: tAs typical range for peanut butter: 6.5–21.4 ppb. iAs fraction: uncertain (0–41% of tAs, n = 3 speciated). The app should avoid conflating total As with inorganic As for peanut-butter-containing products.

Verification notes

Merge-enhanced 2026-05-28 from the v2.0 ingest-next-manual-fetch-pdf workflow. Defects corrected:

• raw_handle was manual-fetch-kimi (generic literal); corrected to canonical MFK_arsenic-speciation-in-commercially-available-peanu to match the manual-fetch handle convention.
• raw_path filename was truncated (...peanut butter spread.pdf) and missing the trailing space in Manual Fetch Kimi; corrected to the on-disk filename Arsenic speciation in commercially available peanut butter spread by IC-ICP-MS.pdf under raw/Manual Fetch Kimi /condiment_papers/05_PB_Vanilla_Spices/.
• Added access_url (DOI link) and raw_sha256 (SHA-256 of the on-disk PDF).
• Added an explicit sample-counting note in sample_population to disambiguate the abstract’s “six peanut butter spread samples” from the eight values reported in Table 3 (six spreads + one whole-peanut + one SRM).
• Reframed SRM 2387 speciation in Key numbers: prior version labelled the first-replicate values (DMA 8.9, MMA 5.5) as “average”, which is incorrect — those are first-replicate values; the abstract’s averages across both replicates are DMA 8.3 / MMA 4.7. Both replicate values are now shown with the abstract averages alongside.
• Tightened the opening prose to reflect that inorganic-As species were detected in only one of the three speciated spreads, and added Table-reference citations to Key-numbers blocks for auditability.

license was previously unknown; updated to subscription (RSC J. Anal. At. Spectrom. is a paywalled journal; the manuscript copy in raw/ is for internal evidence-grading use only). No structural taxonomy slugs changed; routing target set (products/peanut-butter, ingredients/peanut-butter, ingredients/peanuts, metals/arsenic) was already correct.

Audit subagent (2026-05-28) flagged the Certification/Implications phrase “low relative to rice-based ingredients” as a cross-source comparative claim the paper itself does not make. Verified against the source — the Hovanec 2004 manuscript reports peanut-butter values in isolation and does not benchmark them against rice. The rice comparison was synthesis-pass content authored into the source page; removed per Part 5 source-page discipline (cross-source synthesis belongs in the Part 9 workflow, not on individual source pages). All other audit checks (1 numerical fidelity, 2 slug vocabulary, 3 speciation/methods, 4 Part 12 brand firewall) returned clean.

Wiki pages updated on ingest

• peanut-butter
• peanuts
• peanut-butter
• arsenic

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.