Nelson 2023 — Cadmium and other metals in Washington children’s jewelry, 2018 follow-up (WSDOE 23-03-004)

This Washington State Department of Ecology Environmental Assessment Program publication (Nelson 2023, Publication 23-03-004) is a 2018 follow-up to the agency’s 2015 children’s jewelry survey (Sekerak 2016, Publication 16-03-007), conducted under the Children’s Safe Products Act (CSPA, Chapter 70.240 RCW) to assess continued compliance with Washington’s 40 ppm cadmium and 90 ppm lead restriction limits for children’s products. Ecology purchased 78 children’s jewelry items and apparel-with-jewelry sets in October-December 2018, separated them into 555 components, prioritised 38 components via XRF screening, and submitted them to the Manchester Environmental Laboratory for ICP-MS analysis (EPA 6020B following EPA 3052 microwave digestion) of antimony, arsenic, cobalt, mercury, cadmium, and lead. Of the 38 component samples, 37 contained at least one metal above the laboratory reporting limit; 11 (29%) exceeded the 40 ppm CSPA cadmium limit, 7 exceeded the 90 ppm CSPA lead limit, and 4 exceeded the 100 ppm CHCC reporting threshold for antimony. The highest individual values were cadmium at 966,000 ppm (96.6% Cd by mass) in a silver-coloured metal pendant composite from a necklace sold with a lace top and leggings, and lead at 510,000 ppm (51.0% Pb by mass) in the metal back of rhinestone decorations from a bracelet sold with a coral dress.

Key numbers

Sample frame

• Products purchased: 78 (Apparel with Jewelry n=38; Necklaces n=16; Bracelets n=13; Earrings n=7; Jewelry Variety Packs n=6; Watches n=1) (Table 1, p. 4).
• Total individual components yielded by disassembly: 555 (p. 4).
• Components submitted to MEL for ICP-MS analysis: 38, drawn from 33 of the 78 products (p. 4).
• Sample-matrix distribution of the 38 lab samples: metal-matrix n=32; plastic-matrix n=4; fabric-matrix n=2 (p. 4).
• Composite samples (n=2): 1903043-25 combined silver-coloured metal pendant front JC-8-7-6 with pendant back JC-8-7-7 (from a necklace sold with a lace top and leggings); 1903043-28 combined silver-coloured microphone earring decoration CP-7-2-9 with silver-coloured star earring decoration CP-7-2-17 (from the same earring set) (p. 4).
• Glass samples not analysed (n=2): 1903043-02 and 1903043-06 — MEL’s digestion procedure does not use hydrofluoric acid, and Ecology did not select a contract laboratory for these two glass components (p. 5).
• Retailer sourcing: 11 south-Puget-Sound retail stores and 6 online retailers (17 retailers total) (p. 3).
• Purchase window: October, November, and December 2018 (p. 3).
• Analytical processing date: sample coolers delivered to MEL on 1 April 2019 (40 samples) and 3 April 2019 (1 sample); chain-of-custody updates on 13 May 2019 (p. 5).
• Samples with at least one detected metal: 37 of 38 (p. 6).

Detection frequencies and detected-value ranges (Table 2, p. 7)

Reporting limit (RL) was 1 ppm for most samples; raised to 9.80 ppm for 1903043-29 and to 9.33 ppm for 1903043-30 due to dilution to mitigate matrix interferences, and to 2.40 ppm for 1903043-11 due to limited sample mass (p. 6).

Distribution of detections by product type (p. 6-7)

• Lead detected in 36/38 components (95%): 17 from apparel-with-jewelry, 7 from bracelets, 6 from necklaces, 5 from earrings, 1 from a jewelry variety pack. Highest detections in metal components of necklaces and bracelets.
• Cadmium detected in 30/38 components (79%): 14 from apparel-with-jewelry, 5 from bracelets, 5 from necklaces, 5 from earrings, 1 from a jewelry variety pack.
• Antimony detected in 16/38 components: 7 from apparel-with-jewelry, 4 from bracelets, 3 from necklaces, 2 from earrings.
• Arsenic detected in 12/38 components; no per-product-type breakdown reported.
• Cobalt detected in 16/38 components; no per-product-type breakdown reported.
• Mercury not detected in any component.

Exceedances of CSPA limits (Table 3, p. 8; Compliance Summary, p. 9)

CSPA limits used as compliance reference: cadmium ≤ 40 ppm, lead ≤ 90 ppm, CHCC reporting threshold for any chemical of high concern to children = 100 ppm.

Eleven cadmium-exceedance samples (all metal-matrix): 1903043-05 (96.8 ppm), 1903043-07 (75.9 ppm), 1903043-15 (51.1 ppm), 1903043-18 (63.1 ppm), 1903043-26 (109 ppm), 1903043-27 (94.8 ppm), 1903043-28 (79.7 ppm, J-qualified, composite), 1903043-35 (72 ppm), 1903043-25 (966,000 ppm = 96.6%, composite), 1903043-32 (567,000 ppm = 56.7%), 1903043-34 (588,000 ppm = 58.8%) (Table 3, p. 8).

Seven lead-exceedance samples: 1903043-34 (1,020 ppm metal-matrix), 1903043-12 (94,100 ppm = 9.41%, metal), 1903043-29 (173,000 ppm = 17.3%, metal), 1903043-30 (510,000 ppm = 51.0%, metal, J-qualified), 1903043-20 (1,280 ppm, J-qualified, fabric-matrix), 1903043-22 (314 ppm, J-qualified, plastic-matrix), 1903043-36 (402 ppm, J-qualified, plastic-matrix) (Table 3, p. 8).

Four antimony-exceedance samples: 1903043-34 (122 ppm metal-matrix), 1903043-22 (2,800 ppm J-qualified plastic), 1903043-36 (1,670 ppm J-qualified plastic), 1903043-03 (3,720 ppm J-qualified metal-matrix) (Table 3, p. 8).

Extreme high-value component findings (p. 9)

• Three components with cadmium above 500,000 ppm (50% Cd by mass):
  • 1903043-25 (composite of pendant front JC-8-7-6 and pendant back JC-8-7-7): silver-coloured metal pendant from a necklace sold with a lace top and leggings — 966,000 ppm (96.6% Cd).
  • 1903043-34 (SK-17-7-7): metal back of a charm from a necklace of a blue floral-print multi-piece set — 588,000 ppm (58.8% Cd).
  • 1903043-32 (SK-17-5-4): metal horse charm from the necklace of a pink-unicorn multi-piece set — 567,000 ppm (56.7% Cd).
• Three components with lead above 9,000 ppm:
  • 1903043-30 (JC-7-3-5): metal back of rhinestone decorations from a bracelet sold with a coral dress — 510,000 ppm (51.0% Pb), J-qualified (incomplete digestion).
  • 1903043-29 (JC-7-3-3): metal chain from the bracelet sold with the same coral dress — 173,000 ppm (17.3% Pb).
  • 1903043-12 (AM-30-1-5): gold-coloured metal pendant cube from a necklace sold with a chevron dress — 94,100 ppm (9.41% Pb).

Cross-reference to the 2015 study (Sekerak 2016) (p. 3)

• Lead detection frequency, 2015 study: 97% of submitted samples (38 jewelry items).
• Cadmium detection in 2015 study necklaces sold with children’s dresses: range 39.7% to 98.4% Cd by mass.
• Highest lead concentration in 2015 study: 5% Pb (50,000 ppm) in a necklace component sold with a child’s dress.
• 2015 study items with at least one target metal above 100 ppm CHCC threshold: 13 of 38 (34%).

Aggregated CHCC-threshold summary, 2018 study

Of the 38 components in the 2018 follow-up, the highest-priority finding (per CSPA’s per-analyte reporting threshold structure) is the cadmium picture: 11 components (29%) above the 40 ppm restriction, and 3 components at percent-by-mass concentrations (567,000-966,000 ppm). The lead picture is also category-defining: 3 components at percent-by-mass concentrations (94,100-510,000 ppm), all from metal components of jewelry sold as apparel accessories (the AM and JC component-code series correspond to necklaces and bracelets sold with children’s dresses). Antimony, arsenic, and cobalt detections are at orders of magnitude below the CSPA thresholds. Mercury was not detected. The 2018 finding pattern is consistent with the 2015 study’s category-level finding that the highest toxic-metal concentrations in children’s jewelry occur in metal jewelry components sold as accessories with children’s apparel.

Methods (brief)

Component samples were prioritised for laboratory analysis using X-Ray Fluorescence (XRF) screening, with highest priority assigned to components showing screening evidence of lead or cadmium. The 38 selected components were digested by microwave digestion following EPA Method 3052 and analysed for antimony, arsenic, cobalt, mercury, cadmium, and lead by inductively coupled plasma mass spectrometry (ICP-MS) following EPA Method 6020B at the Washington State Department of Ecology Manchester Environmental Laboratory (MEL). Samples were digested in three batches: one all-metal-matrix, one all-plastic-matrix, and one mixed metal-and-fabric. Two glass-matrix samples (1903043-02, 1903043-06) were not digested because MEL does not use hydrofluoric acid as part of its digestion process; no contract laboratory was engaged for these components. Quality control followed the project QAPP addendum (Sekerak 2018); CCB and CCV checks were outside acceptance limits for cadmium on five samples (1903043-12, -20, -03 with duplicate, -22, -36), and the corresponding cadmium results were qualified as estimates. Matrix-spike duplicate RPDs for antimony were above QC limits for sample 1903043-03 (antimony result qualified “J”), and cobalt and cadmium RPDs were above QC limits for sample 1903043-28 (those results qualified “J”). Matrix-spike recoveries for antimony in 1903043-03 and for lead in 1903043-20 could not be evaluated because the standard spiking level was insufficient for the elevated source-sample concentrations. Reporting limits were raised from the nominal 1 ppm to 9.80 ppm (1903043-29) and 9.33 ppm (1903043-30) due to dilution to mitigate matrix interference, and to 2.40 ppm (1903043-11) due to limited sample mass. Seven samples (1903043-11, -30 metal; -20 fabric; -01, -03, -22, -36 plastic) did not completely digest; the associated results were qualified “J” and may be biased low. Chain of custody was maintained throughout collection, processing, and lab transfer, with one field-identification discrepancy (1903043-27) reconciled and documented in the case narrative. Two component identifications were composited prior to lab submission to reach the required digestion mass (composites JC-8-7-7 + JC-8-7-6 and CP-7-2-9 + CP-7-2-17). Full project plan: Sekerak 2018, Addendum to Quality Assurance Project Plan: Product Testing Program, Version 1.0 — Product Testing Follow-up Study 2018-2019, Publication 18-03-113.

Implications

• Children’s-product compliance baseline (WA state, CSPA). This study is the agency follow-up to the 2015 baseline (Sekerak 2016) for Washington State’s CSPA compliance-monitoring programme on children’s jewelry. The 2018 detection-frequency picture (lead in 95% of 38 prioritised components; cadmium in 79%) and the per-mass extreme-value findings (cadmium up to 96.6% by mass; lead up to 51.0% by mass) are the agency-reported continued non-compliance signal that Ecology submitted to the CSPA compliance lead for further assessment. The study does not propose a new threshold; it documents continued exceedance of the 40 ppm cadmium and 90 ppm lead CSPA restrictions in children’s jewelry sold in Washington retail and online channels in late 2018.

• HMTc audit implications. For any HMTc certification work that addresses children’s jewelry or jewelry-adjacent accessories sold with children’s apparel, this study provides the agency-tested 2018 baseline distribution of Pb, Cd, Sb, As, Co, and Hg occurrence in the metal, plastic, and fabric matrices of children’s jewelry components. The study’s prioritisation method (XRF screening of all 555 components, ICP-MS confirmation of the 38 highest-XRF prioritised) is a credible category-level worst-case-targeting methodology; the resulting distribution is the upper-tail picture, not a random-sample picture, and HMTc threshold-setting that cites this study should preserve that distinction. The category-level finding — that the most extreme Pb/Cd concentrations occur in metal jewelry components sold as accessories with children’s apparel (the AM and JC component-code series in this study) — is the actionable contamination-platform finding for downstream Step 0 Lock work on children’s jewelry as a category.

• App. Not directly relevant to ingredient contamination_profile data because no food-matrix occurrence values are reported. Relevant to a future children’s-product screening surface that explains why metal jewelry components — particularly jewelry sold as apparel accessories — can be a high-Pb, high-Cd exposure source by ingestion (small parts that can be mouthed or swallowed) or by dermal/saliva contact.

• Courses. Worked example for the regulatory-compliance module of how a state agency uses XRF screening to prioritise component samples for confirmatory laboratory analysis under a per-product-class compliance-monitoring statute (CSPA Chapter 70.240 RCW). Also a primary-source illustration of the gap between (i) the substance-level restriction limit (40 ppm Cd, 90 ppm Pb) and (ii) the contaminant-reporting threshold (100 ppm CHCC) that triggers manufacturer reporting under Chapter 173-334 WAC.

• Microbiome. Not applicable. No biological matrix and no exposure-pathway measurements relevant to gut microbiome interaction.

Wiki pages this source may touch

• lead
• cadmium
• antimony
• arsenic-total
• cobalt
• mercury-total
• childrens-jewelry

Verification notes

• Source identification. Nelson, Ken. 2023. Cadmium and Other Metals in Children’s Jewelry 2018, Follow-up Study. Publication 23-03-004. Washington State Department of Ecology, Olympia, WA. Nine-page agency study issued by Ecology’s Environmental Assessment Program in January 2023. Available at https://apps.ecology.wa.gov/publications/SummaryPages/2303004.html.
• Author attribution. Single author Ken Nelson, with the agency’s Environmental Assessment Program. The 2015 predecessor study was authored by S. Sekerak (Publication 16-03-007); the 2018 QAPP addendum was authored by Sekerak (Publication 18-03-113). Cited per the agency’s suggested citation as Nelson 2023.
• DOI. No DOI assigned; Washington State agency publications are not normally DOI-registered. doi: null with no_doi_assigned: true and access_url populated.
• License. Washington State agency publication; public-domain under state-government work convention. The publication explicitly notes “Any use of product or firm names in this publication is for descriptive purposes only and does not imply endorsement by the author or the Department of Ecology” (p. 2).
• Tier rationale. A-tier per Part 13: government regulatory-compliance study from a state environmental agency operating under statutory authority (CSPA Ch. 70.240 RCW), conducted by an accredited state laboratory (MEL) under a formal QAPP. Matches the existing handling of state and federal agency product-testing publications (padoh2023-pahan-727-cpsc-lead-cup-recall, cpsc1997-pvc-children-products).
• Source-type rationale. source_type: regulatory: the document is a state agency compliance-monitoring publication issued under CSPA’s compliance-assessment authority, with lab results “submitted to Ecology’s CSPA compliance lead for further assessment” (Abstract, p. 1). Matches the corpus convention for state-agency CSPA/Prop-65/CPSC compliance studies.
• Frontmatter metals: [Pb, Cd, Sb, tAs, Co, tHg] reflects the six analytes the study reports (lead, cadmium, antimony, arsenic, cobalt, mercury). Arsenic abbreviation is tAs (total) — the study does not speciate inorganic from total arsenic; ICP-MS following EPA 6020B as run reports total elemental arsenic. Mercury abbreviation is tHg (total) — the study does not separate methylmercury from total mercury; ICP-MS under EPA 6020B reports total mercury. Sb is total antimony. Co is total cobalt. Pb and Cd carry no speciation question at the conventional ICP-MS reporting level.
• Frontmatter ingredients: [] is correct — no food ingredients involved. This is a consumer-product compliance-monitoring study of jewelry items.
• Frontmatter matrices: [] is correct — no food or biological matrices. The lab samples are metal, plastic, fabric, and glass component-of-jewelry matrices; these are hard-goods consumer-product matrices, not the food/biomonitoring controlled-vocabulary matrices the routing layer routes on.
• Frontmatter products: ["[[products/childrens-jewelry]]"] uses the broad children’s-jewelry umbrella slug. The slug childrens-jewelry does not exist in the 2026-05-18 taxonomy snapshot as a populated product page; it is created as a provisional scaffold by tools/autonomy/create-provisional-product-scaffold.mjs in the same ingest commit as this source page, per CLAUDE.md Part 10 (“product pages are auto-created as routing destinations whenever a source declares a slug with no matching page”). HMTc category assignment, clean/contaminated pairing, and threshold-setting on the children’s-jewelry product category do not derive from this page until a Step 0 Lock is authored (separate workflow, Karen-owned). The single broad slug is chosen rather than separate sub-form slugs (necklaces, bracelets, earrings, watches, jewelry-variety-packs, apparel-with-jewelry) because: (i) the study reports aggregate detection statistics across all jewelry types together (Table 2) rather than by-jewelry-type detection summary statistics; (ii) per the Phase 1 broad-scope rule, frontmatter is not over-narrowed to specific sub-categories the source does not separately analyse statistically; (iii) the by-type detection counts in the body text (e.g., “17 from apparel-with-jewelry, 7 from bracelets, 6 from necklaces”) are present as descriptive distribution but do not constitute per-sub-type occurrence summary statistics suitable for routing as direct-evidence to multiple separate product pages.
• Frontmatter jurisdictions: [US] — Washington state agency publication issued under Washington state statute. Convention in the corpus is country-level codes; the Washington-specific state context is captured in the page body and in the publication: field rather than in a state-level jurisdiction code.
• Brand-firewall (Part 12, strict reading locked 2026-05-17). The source identifies products by an internal Ecology field-collection code (the AM/CL/CP/DT/FM/JC/NR/SK/WM letter prefix designates the collector / retail location coding scheme, not the brand) and by descriptive component descriptors (“metal horse charm”, “gold-coloured metal pendant cube”). The source mentions one branded product name in the methods section: a “Word Wear Personalized Jewelry Set (AM-30-6)” referenced once on p. 5 as part of a chain-of-custody discussion about a sample-matrix code change (plastic to glass to fabric replacement chain). That brand name is not transcribed into this wiki page; the chain-of-custody event is summarised in the Methods section without naming the brand, consistent with the Part 12 strict reading that the wiki defaults to no brand attribution. The descriptive component descriptors used in the source (jewelry sold with “a chevron dress”, “a coral dress”, “a lace top with leggings”, “a pink unicorn multi-piece set”, “a blue floral print multi-piece set”, “a zig-zag print dress”) are apparel-form descriptors, not brand names — they describe the print/style of the children’s dress with which the jewelry was sold, and are preserved as the product-form descriptor on this wiki page (the equivalent of “the preterm formula” rather than naming the brand). The Ecology component-code series (AM, JC, SK, etc.) are internal sample-identification codes only, not brand identifiers.
• Scientific-method vendor names (Part 12 Exception 2). Methods section names the analytical method (ICP-MS following EPA 6020B), the digestion method (microwave digestion following EPA 3052), the lab (Washington State Department of Ecology Manchester Environmental Laboratory, MEL), and the agency programme (Environmental Assessment Program). These are scientific-method identifications, not contamination-value brand attributions, and are kept per the locked Exception 2 reading.
• Wiki/HMTc firewall (Part 2). The Implications section reports the agency’s documented finding and the category-level distribution as expressed in the source. It does not propose HMTc threshold values, does not compare this study to other literature, does not soften or strengthen the reported values. The HMTc audit implications bullet describes what the study contributes to threshold work (an upper-tail-targeted distribution from XRF prioritisation) without proposing a threshold, on the allowed side of the Part 2 boundary per audit-prompt.md.
• Speciation flag. As above: arsenic is total (tAs) and mercury is total (tHg) because the source does not speciate; no MeHg, iAs, iHg, or Cr-VI claims are made. Lead, cadmium, antimony, and cobalt carry no speciation question at the EPA 6020B ICP-MS level.
• Regulation references. The source cites Washington’s Children’s Safe Products Act (CSPA, Ch. 70.240 RCW) restriction limits (40 ppm Cd, 90 ppm Pb) and the 100 ppm CHCC reporting threshold under Ch. 173-334 WAC. Neither of these Washington state regulations currently has a dedicated wiki/regulations/ page in the 2026-05-18 taxonomy snapshot; the page references them by name and statutory citation in the body and Compliance Summary discussion rather than wikilinking to non-existent regulation slugs. Not created speculatively from this single source.
• Folder context vs paper scope. The PDF lives under _extracted_infantcontact_02_Teethers_Pacifiers/02_Teethers_Pacifiers/ in the Kimi corruption-issue raw tree, but the document is not a teether/pacifier study — it is a children’s jewelry compliance follow-up. The folder naming reflects the Kimi-agent’s batch-organisation scheme during the May 21 corruption-recovery extraction, not the paper’s actual content. The product slug is childrens-jewelry, not pacifiers-and-sucking-teething-aids.
• Number transcription note. Page 6 paragraph 4 of the source contains a transcription typo (“510,00 ppm (51.0%)”); the correct value confirmed against Table 3 (p. 8) and Compliance Summary bullet on p. 9 is 510,000 ppm (51.0%) for sample 1903043-30 (JC-7-3-5). The wiki page uses the correct 510,000 figure.
• Sample composite annotation. Two samples (1903043-25 and 1903043-28) are pendant-front-plus-pendant-back and earring-decoration composites respectively, combined before submission to reach required digestion mass. The headline 966,000 ppm Cd finding is therefore a composite-sample measurement of two components from the same pendant, not a single-component measurement. The wiki page preserves the composite annotation throughout (Key numbers, sample-population, Table 3 transcription, extreme-value bullets).
• Near-duplicates. None identified in the corpus at this writing. The 2015 predecessor study (Sekerak 2016, Publication 16-03-007) is referenced in this paper but is not yet ingested. The 2018 QAPP addendum (Sekerak 2018, Publication 18-03-113) is referenced as the project plan but is a procedural document, not an occurrence study. Neither is ingested in this commit.
• Raw integrity. raw_sha256 = 089e91df9657b15c537ea05361c79ea21443926bdb61292eee1058e6fbfd4055 verified by shasum -a 256 against the file at raw_path.

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.