Munarso et al. 2024 — Chemical quality and production dynamics of Indonesian vanilla

Survey-and-laboratory study of cured Indonesian vanilla beans (Vanilla planifolia Andrews) sampled May–August 2023 across seven provinces, conducted by Indonesia’s National Research and Innovation Agency (BRIN) with the National Food Agency (NFA) to inform a Codex Alimentarius draft standard for vanilla. The paper’s primary focus is vanillin content and physical/chemical quality across four export grades, plus production constraints (theft-driven premature harvest, curing variation between farmers and exporters). Heavy-metal data (Pb and Cd only) appear in a single contaminant table (Table 6) framed by the authors as evidence supporting a “clean-green-safety” positioning of Indonesian vanilla in the export market.

Key numbers

Heavy-metal concentrations in cured vanilla beans, by quality grade and source (Table 6, mg/kg, mean ± SD):

• Pb across the seven grade × source rows: row means 0.03–0.15 mg/kg; standard deviations are larger than the means in three rows (Grade A exporter 0.09 ± 0.15; Grade B exporter 0.13 ± 0.19; Broken Grade D farmer 0.15 ± 0.22), indicating individual samples in those rows ran substantially higher than the row means.
• Cd across the seven rows: row means 0.01–0.06 mg/kg; SDs exceed the means in the Grade A exporter (0.04 ± 0.110) and Grade B exporter (0.06 ± 0.130) rows, indicating the upper-end individual samples in those rows are roughly an order of magnitude above the row means.
• No analytical speciation reported. iAs, tAs, tHg, MeHg, Ni, Al, Cr, Cr-VI, Sn, Sb, and U were not measured.

Chemical quality (Table 3, mean ± SD; parenthesised values are Indonesian SNI minimum/maximum):

Physical quality (Table 4): pod length ranged 11.60 ± 4.78 cm (Broken Grade D, SNI >8) to 19.71 ± 0.81 cm (Gourmet/Grade A, SNI >11); foreign-object, shriveled-immature, and mold counts were 0 across all four grades for the sampled set.

Sampling: 68 individuals interviewed (8 exporters + 60 farmers) across 7 provinces. Vanillin tested by spectrophotometric method (ISO 5565-2:1999); water and ash by SNI 01-0010-2002 (water/ash) and SNI 01-2891-1992 (ash insoluble in acid). The paper does NOT state which analytical method was used for the Pb/Cd determinations in Table 6.

Methods (brief)

Survey-and-sample design combining purposive sampling of long-experienced vanilla exporters (>10 years of export experience, own plantation, government-recognised) with snowball sampling of four farmers per exporter, conducted May–August 2023 in North Sumatra, Central Java, East Java, NTT, NTB, Bali, and North Sulawesi. A structured questionnaire was administered to 68 individuals (8 exporters and 60 farmers). For chemical-quality testing of Tables 3/4 (water, vanillin, ash, ash insoluble in acid), n=4 samples per quality class were measured. Statistical analysis used one-way ANOVA with Duncan’s post-hoc test in IBM SPSS 26; significance at p<0.05. Vanilla quality parameters were measured against Indonesian National Standards (SNI) per the method matrix in Table 2 (water and color: SNI 01-0010-2002; ash insoluble in acid: SNI 01-2891-1992; vanillin: spectrophotometric per ISO 5565-2:1999).

The Pb and Cd values in Table 6 are described only as “main chemical contaminants” measured in a “certified laboratory”; no instrument (AAS / ICP-OES / ICP-MS), no digestion procedure, no LOD/LOQ, no reference material, no quality-control summary, and no per-row sample size are reported for the heavy-metal determinations. Recovery/precision data for Pb and Cd are absent. The paper does not separate inorganic from total arsenic or methylmercury from total mercury because it does not measure arsenic or mercury at all.

Explicit limitations: (1) no analytical method stated for Pb/Cd; (2) only Pb and Cd measured among heavy metals; (3) per-row sample size for Table 6 not stated and standard deviations larger than means in several rows indicate substantial within-row variability; (4) single-country (Indonesia) source data; (5) Table 6 values are reported descriptively without inferential testing.

Implications

Certification: Provides preliminary Indonesian occurrence data for Pb and Cd in cured vanilla beans across export grades. The absence of a reported analytical method, LOD/LOQ, and per-row sample size limits how the Table 6 values can be used as occurrence inputs without further methodological assumptions. The absence of iAs, tAs, tHg, MeHg, Ni, Al, Cr, Cr-VI, Sn, Sb, and U measurements leaves the other HMI/HMT&C analytes as data gaps for vanilla specifically.

Courses: Useful as a case study in spice-export quality control, supply-chain variance (farmer-versus-exporter curing practices), and the Codex standards-development pipeline (the paper was funded by Indonesia’s NFA explicitly to feed Codex CCSCH discussions on a vanilla standard).

App: Pb and Cd values for cured vanilla beans (Indonesia, 2023); arsenic, mercury, nickel, aluminium, chromium, and tin remain data gaps for this matrix from this source.

Verification notes

• 2026-05-28 (Claude Opus 4.7, v2.0 manual-fetch merge-enhance — Phase 3 audit application): fresh-context audit subagent (Phase 2) returned PROMOTE with two cosmetic ⚠️ items. Both verified against the PDF and applied: (1) the Table 6 reproduction’s “Whole Grade A (Gourmet)” row label was simplified to “Whole Grade A” to match the source’s Table 6 column verbatim (Tables 3 and 4 use the longer “Gourmet/Whole Grade A Class 1A” descriptor, but Table 6’s own label is just “Whole Grade A”); (2) the “Vanilla is not currently a standalone ingredient page in the wiki” parenthetical was removed from the Implications > App paragraph (wiki-state observations belong in Verification notes, not in Implications, per Part 2 firewall hygiene). The remaining ⚠️ items (Duncan grouping superscripts dropped from Table 3 reproductions; ISO 5565-2:1999 citation form refined from Table 2’s “Spectrophotometric methods / ISO (1999)” cell to the full references-list designator) were verified as faithful-to-source and left unchanged — the auditor itself classified them as cosmetic.
• 2026-05-28 (Claude Opus 4.7, v2.0 manual-fetch merge-enhance — Phase 1): page was originally created 2026-05-14 in the kimi-cond05 batch ingest. Phase 1 enhancements applied:
  • Set per-PDF raw_handle: MFK_from-bean-to-market-exploring-the-chemical-and-pro (was the generic batch handle manual-fetch-kimi).
  • Restored full raw_path with the trailing-space folder name (“Manual Fetch Kimi ”) matching the on-disk filename “From bean to market_ exploring the chemical and production dynamics of high-quality Indonesian vanil.pdf” (prior path truncated the filename mid-word at “dynami.pdf”).
  • Added raw_sha256 (6c9389d…1dd93), access_url (https://doi.org/10.3389/fsufs.2024.1425656), and tier_rationale explaining the B-tier placement.
  • Populated products: ["[[products/spices]]"] (was empty, which surfaced this source in routing_malformed.csv as “Source is missing optional routing-input fields”); the spices product page exists and is the broadest accurate routing destination for a vanilla-bean paper.
  • Replaced the invented [[ingredients/vanilla]] slug with [[ingredients/herbs-and-spices]]: vanilla is not currently an ingredient page in the taxonomy snapshot, and per the hard constraint against creating new ingredient pages in this session, the routing default is the broadest accurate slug that exists (herbs-and-spices). If vanilla later promotes to its own page, the routing layer will fan out.
  • Replaced the non-canonical matrix vanilla-bean-cured with the canonical matrix spices (matrices is a bare-string vocabulary; the system-prompt matrices list does not include vanilla, and other condiment_papers/05 sources route under spices).
  • Added sampling_locations and sampling_year_range to capture the seven-province scope and the May–August 2023 sampling window now that the source has been re-read.
  • Set sample_n: null (was 28, which is an interpolation rather than an explicit count); the paper states n=4 per quality class for Tables 3/4 chemical-quality testing but does NOT state the per-row sample size for Table 6 heavy-metal data, so the prior 4 × 7 = 28 figure was inferred. The narrative now records what the source actually says.
  • Expanded Key numbers to include the full Table 3 chemical-quality matrix and the Table 4 physical-quality summary alongside the Table 6 contaminant table, since this paper’s metals-only contribution is bounded and the broader quality framing is what the rest of the paper is about. Standard deviations that exceed row means are now called out explicitly to flag the within-row spread.
  • Expanded Methods (brief) to document that no instrument, no LOD/LOQ, no digestion, no QC, and no per-row n is given for Pb/Cd; this is a Cochrane-style limitation note rather than a synthesis claim.
  • Rewrote the Implications section to remove the prior cross-source comparison to “the Codex general ML for spices (Pb 2.0 mg/kg, Codex CXS 193-1995)” — that comparison is a synthesis claim (Part 2 wiki/HMT&C-firewall hygiene); it belongs on the relevant regulation page or in a downstream synthesis pass, not in a single-source page. The underlying point — that this paper does not measure most HMI/HMT&C analytes — is preserved as a data-gap statement. Also removed the prior “consistent with vanilla being a relatively low-contamination spice” line for the same reason.
  • Replaced the legacy ## Wiki pages updated on ingest heading with the current template ## Wiki pages this source may touch (the legacy heading is flagged by the audit checklist in docs/gpt-collaboration/verification-checklist.md).
• DOI, title, authors, year, publication, license, evidence_tier, and metals list (Pb, Cd) verified against page 1 (citation block) and pages 8–9 (Table 6 caption and conclusion); unchanged.
• Table 6 Pb and Cd values reproduced exactly from the source; no rounding, no unit conversion (mg/kg in source = mg/kg in wiki).

Wiki pages this source may touch

• herbs-and-spices
• spices
• lead
• cadmium

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.