Heavy Metal Index

Experimental study for inorganic and organic profiling of toy makeup products: Estimating the potential threat to child health

14 min read

Source access

Publication
Environmental Science and Pollution Research 31: 33975–33992
Year
2024
Authors
Mercan S; Kilic MD; Zengin S; Yayla M
Source type
peer reviewed
License
CC BY 4.0

Mercan et al. 2024 — 14-element inorganic and allergen-fragrance profile of toy makeup products from Istanbul, with health-risk modelling

This open-access Environmental Science and Pollution Research paper quantifies 14 inorganic elements (Al, As, Ba, Cd, Co, Cr, Cu, Hg, Mn, Ni, Se, Sb, Pb, Zn) by ICP-MS in 63 toy makeup products (eye shadow, nail polish, lipstick-type cosmetic toys) purchased from Istanbul toy stores and computes systemic exposure dosage (SED), margin of safety (MoS), lifetime cancer risk (LCR), hazard quotient (HQ), and hazard index (HI) per SCCS 2012 dermal-exposure modelling using a 5-year-old child receptor. The qualitative organic profile of the same products is screened by GC–MS for allergen fragrance. Of 63 toy makeup samples, 57 (90.48%) exceeded at least one regulatory permissible limit for a toxic element, with descending exceedance order Ni > Cr > Co > Pb > Sb > Cd > As > Hg. The paper concludes that toy makeup products marketed in Istanbul toy stores are predominantly unsafe for children and calls for regular monitoring and stricter enforcement.

Key numbers

All inorganic concentrations are total elemental in the whole toy-makeup product (µg/g of product as placed on market). Samples were oven-dried at 37 °C and homogenized before microwave acid digestion (CEM Mars 5, concentrated HNO₃ 65%) followed by ICP-MS (Thermo Scientific X Series-II + CETAC ASX 520 auto-sampler), n=3 replicates per sample.

Method performance (p. 33978–33979)

  • Linearity: 11-point calibration over 0.1–150 ng/mL for Al, As, Ba, Cd, Co, Cr, Cu, Mn, Ni, Se, Sb, Pb; 12-point calibration over 1–400 ng/mL for Zn; 0.5–20 ng/mL for Hg. r² ≥ 0.999.
  • LOD: 0.03–0.30 ng/mL for trace elements (Ba, Co, Cr, Cu, Mn, Se, Zn); 0.03–0.39 ng/mL for toxic elements (Al, As, Cd, Hg, Ni, Sb, Pb).
  • LOQ: 0.11–0.99 ng/mL (trace); 0.10–1.29 ng/mL (toxic).
  • CRMs: BCR-700 organic-rich soil and CRM 7002 light sandy soil. Recovery 71.77–112.21%.
  • Standard-addition spike recovery (no certified makeup CRM): 89.00% (Ni) to 110.17% (Ba) at 10 ng/mL; 93.00% (Ni) to 106.17% (Sb) at 20 ng/mL; 110.00% (Hg) at 5 ng/mL; 89.00% (Hg) at 10 ng/mL.
  • RSD across n=3 spike replicates < 16.64%.
  • Internal standards: Ga and In (20 ng/mL each), added to all calibrators and samples.

Inorganic concentration distributions across all 63 samples (Table 1 across 10 brands; ranges across brand means and brand min–max)

The source presents Table 1 as min–max, mean, median, and SD per brand for each of the 14 elements rather than a single 63-sample distribution. The values below summarize the brand-mean ranges and the across-brand absolute min–max in µg/g of product.

ElementAcross-brand mean: lowest to highest (µg/g)Single-sample min–max across all brands (µg/g)Limit-exceedance count and rank
Al1121 – 25,117116 – 37,970Mean limit-exceedance reported for 57.15% of samples (combined Al/As/Cd/Co/Hg/Mn/Sb/Zn MoS exceedances); Al-specific SED > TDI in 54.55% of brand means (n=6 of 11 brand-mean entries reported in source) and Al worst-case-95th-percentile SED > TDI in 7 of 10 brands
Cr0.30 – 7.440.23 – 12.53Skin-sensitization-limit (5 mg/kg in cosmetics, per Basketter et al.) exceeded in 17 of 63 samples; LCR > 1×10⁻⁶ in 100% (n=63)
Mn0.45 – 151.24<LOQ – 258.64MoS-exceedance lever (n=53 brand-mean entries flagged Mn-MoS-unsafe per text); no permissible cosmetic limit cited
Co<LOQ – 2.55<LOQ – 2.88Limit-exceeded in 35 of 63 samples (55.56%); HQ > 1 reported in BRAND-mean for one brand
Ni<LOQ – 29.17<LOQ – 40.15Permissible-cosmetic-limit exceedance 77.78% (n=49) — Ni was the most prevalent limit-exceeder of the eight regulated toxic elements per p. 33981; skin-sensitization-limit (>5 mg/kg, Basketter et al.) exceeded in 6 of 63 samples (9.52%, n=6); LCR > 1×10⁻⁶ in 79.37% (n=50)
Cu<LOQ – 67.30<LOQ – 164.68Single-sample max 164.68 µg/g; no permissible limit cited
Zn<LOQ – 5762<LOQ – 6227Mean limit-exceedance reported under combined Zn/Al/Mn no-limit umbrella; Zn MoS unsafe in 32 brand-means
As<LOQ – 13.03<LOQ – 27.68Single-sample max 27.68 µg/g; 85.71% (n=54) LCR exceedance
Se<LOQ – 1.47<LOQ – 2.74Below cosmetic-toxic-limit; HQ contribution
Cd<LOQ – 0.33<LOQ – 0.812 brands above the Cd permissible limit; 18.87% (n=10) LCR exceedance among the n=53 samples with Cd > LOQ
Sb<LOQ – 0.78<LOQ – 1.6219.05% (n=12) above the Sb permissible limit
Ba0.89 – 294<LOQ – 1103No cosmetic limit; single-sample maximum 1103 µg/g in one brand’s Ba min-max range
Hg<LOQ – 1.32<LOQ – 4.922 samples exceeded the Hg permissible limit; one brand’s single-sample maximum 4.92 µg/g
Pb<LOQ – 81.71<LOQ – 171.20Single-sample maximum 171.20 µg/g (85-fold above the BVL “permissible limit”); LCR-exceedance in 77.78% (n=49) and SED-vs-TDI exceedance in 9.09% (reported as n=8 in source text)

The source body explicitly notes: “the highest Pb concentration was observed … with a concentration of 171.200 µg g⁻¹ which was 85-fold higher than the permissible limit set by BVL” (p. 33979); the highest As detected was 27.68 µg/g (p. 33979); the highest Hg mean reported in any single brand was 1.80 µg/g, with a single-sample maximum of 4.92 µg/g (p. 33979).

Combined exceedance summary across the 63-sample set (Abstract, Conclusion, and pp. 33979–33985)

  • Any-toxic-element exceedance: 57/63 (90.48%) toy makeup samples exceeded the maximum permissible limit for at least one toxic element. Descending order of exceedance frequency across the eight regulated toxic elements: Ni > Cr > Co > Pb > Sb > Cd > As > Hg.
  • MoS < 150 (SCCS pediatric safe threshold): 57.15% of samples were unsafe by MoS for at least one of Al, As, Cd, Co, Hg, Mn, Sb, or Zn. Mn (n=53), Al (n=42), and Zn (n=32) were the most MoS-limit-exceeding elements.
  • LCR > 1×10⁻⁶ (IARC carcinogens Cr, Ni, As, Cd, Pb):
    • Cr: 100% (n=63)
    • Ni: 79.37% (n=50)
    • As: 85.71% (n=54)
    • Pb: 77.78% (n=49)
    • Cd: 18.87% (n=10)
    • Descending LCR order: Cr > As > Ni > Pb > Cd.
  • HQ ≥ 1 (non-cancer hazard): 80% of samples exceeded the HQ safe limit (HQ ≥ 1) for at least one inorganic element.
  • SED > TDI (worst-case 95th-percentile exposure): Al SED > TDI in 7 of 10 brands; Pb SED > TDI in 1 brand (worst-case sample 4.92 µg/g context — Hg also exceeded in BRAND I only).
  • Skin-sensitization risk (> 5 mg/kg in cosmetics, Basketter et al.): Cr 26.98% (n=17), Ni 9.52% (n=6). Co did not exceed this threshold in any sample.

Per-product-class context (p. 33977)

The 63 samples represent eye shadow, nail polish, and lipstick-type cosmetic toys. The paper notes that toy makeup is applied to “the child’s thinnest and weak body parts such as the face (5 mm) and eye (2 mm) and stays on the skin for a long time as adult cosmetics” (p. 33976), motivating the SCCS dermal-exposure parameterization used in the SED calculation.

Health-risk model parameters (p. 33978)

  • Body weight: 18.6 kg (5-year-old child, Turkish Medicine and Medical Devices Agency 2005).
  • Skin surface area / body weight (SSA/BW): 1.5 (5-year-old, Turkish Medicine and Medical Devices Agency 2014).
  • AA (amount applied per day): 0.02 g for eye shadow (SCCS 2012).
  • Frequency (F): 2 applications/day; Retention factor (RF): 1.0 (leave-on); Bioaccessibility (BF): 100%; Conversion factor (CF): 10⁻³.
  • MoS = NOAEL / SED; NOAEL = RfD × UF × MF (UF and MF defaults 1 and 100). Safe threshold MoS ≥ 150 (5-year-old child).
  • LCR = SED × SF, where SF (mg⁻¹ kg d) = 0.5 (Cr), 0.91 (Ni), 1.5 (As), 6.7 (Cd), 0.0085 (Pb), per Lim et al. 2018.
  • HQ = SED / RfD; HI = Σ HQ. Safe HQ < 1.
  • RfDs (mg kg⁻¹ day⁻¹) used: Al 1, Cr 0.0195, Mn 0.00096, Co 0.0003, Ni 0.02, Cu 0.04, Zn 0.3, As 0.0003, Se 0.005, Sb 0.0000125, Ba 0.00006, Hg 14, Pb 0.0003, Cd 0.04. (Shomar & Rashkeev 2021.)
  • TDI values (µg kg⁻¹ d⁻¹, Van Engelen et al. 2008, reproduced in Table 2 footer): Al 7.5×10², Cr 5.00×10², Mn 1.40×10³, Co 1.00×10², Ni 1.00×10², Cu —, Zn —, As 1.00×10², Se 1.00×10², Cd 5.00×10², Sb 6.00×10², Ba 1.00×10³, Hg 2.00×10², Pb 3.60×10³ (units as printed in Table 2).

Worst-case 95th-percentile SED outcomes (Table 3, p. 33984)

The worst-case 95th-percentile SED was computed per Hajrić et al. 2022 (rank-sorted, n × 95% index). Across the 14 elements:

  • Al exceeded TDI in 7 of 10 brand entries.
  • Pb exceeded TDI in 1 brand entry (worst-case 95th-percentile sample).
  • As exceeded TDI in 1 brand entry.
  • Cr, Co, Ni, Cd, Sb, Hg, and the non-toxic essentials remained within worst-case TDI for the parameterized 5-year-old child.

Organic / allergen-fragrance screening (p. 33987)

GC–MS qualitative scan (HP 6890 GC, HP 5975B MSD, HP-5MS column 30 m × 250 µm × 0.25 µm; EI, 70 eV; m/z 50–600). Sample prep: 0.25 ± 0.05 g extracted with 5 mL methanol + 0.5 g anhydrous Na₂SO₄, vortexed and sonicated (50 °C, 10 min, 250 W), centrifuged 5 min at 5000 rpm, filtered (0.22 µm cellulose acetate), reconstituted to 500 µL. Identification by MS library (no quantitation; LOD/LOQ/linearity/recovery not computed for organics).

  • Only 2-phenoxyethanol was qualitatively identified across the 63-sample set as an allergen-fragrance ingredient.
  • 2-Phenoxyethanol is reported in the literature (Birnie & English 2006; Dréno et al. 2019; Kolodziej et al. 2022; Scognamiglio et al. 2012) as an anti-microbial preservative associated with contact dermatitis, mucous-membrane irritation, phototoxicity, and photoallergy. SCCS-listed common fragrance allergens (limonene, lilial, geraniol, etc.) were not detected by the qualitative scan.

Methods (brief)

  • Sample set: n=63 toy makeup products from 10 brands purchased at Istanbul toy stores; 8 brands of Turkish origin, 2 brands of Chinese origin (12 Chinese-import samples).
  • Sample prep (inorganic): oven-dried 37 °C, sieved/homogenized, stored in polyethylene bottles. Microwave acid digestion (CEM Mars 5, with temperature and pressure sensors), 65% HNO₃, three independent replicates per spike level (10 and 20 ng/mL). Digestion vessels acid-soaked (10% HNO₃) before sample prep.
  • ICP-MS (inorganic): Thermo Scientific X Series-II ICP-MS + CETAC ASX 520 auto-sampler. Plasma gas 13 L/min Ar (>99.999% purity, Habas, Turkey), nebulizer 0.87 L/min, RF 1400 W, auxiliary 0.8 L/min, 100 sweeps/replicate, spray chamber 3 °C, sample uptake 45 s, dwell 0.01 s. ISs (Ga + In, 20 ng/mL) in all calibrators and samples. ISs-added blanks and CRMs run every 10 samples for QC and carryover monitoring.
  • Sample prep (organic): 0.25 ± 0.05 g extracted with 5 mL methanol + 0.5 g anhydrous Na₂SO₄ (water removal); vortexed; ultrasonicated 50 °C, 10 min, 250 W; centrifuged 5 min at 5000 rpm; supernatant filtered (0.22 µm cellulose acetate); 200 µL aliquot reconstituted to 500 µL final volume.
  • GC–MS (organic): Agilent HP 6890 GC + HP 5975B MSD + HP 7683B autoinjector; HP-5MS column 30 m × 250 µm × 0.25 µm; helium 1 mL/min; injector 250 °C; split 5:1; transfer line 280 °C; ion source 230 °C; oven 50 °C (2 min) → 265 °C at 15 °C/min, hold 12 min (28 min total); EI 70 eV; full-scan m/z 50–600. MS library-based qualitative identification only.
  • CRMs: BCR-700 (Institute for Reference Materials and Measurements, Belgium); CRM 7002 (Czech Metrology Institute) — both soil CRMs, used because no certified-reference cosmetic material exists.
  • Reagents: Merck Suprapur HNO₃ 65% for digestion; High-Purity Standards 10 mg/mL multi-element + 1000 mg/mL mono-element Hg (Charleston, SC, USA); Absolute Standards (Hamden, CT, USA) Ga and In ISs 1000 mg/mL; Merck Suprapur methanol; Sigma-Aldrich anhydrous Na₂SO₄ (99%); Millipore Direct-Q®3 UV ultrapure water (18.2 MΩ cm).
  • Risk model: SCCS 2012 dermal-exposure model for leave-on cosmetics applied to a 5-year-old child (BW 18.6 kg, SSA/BW 1.5, AA 0.02 g, F 2/day, RF 1.0, BF 100%, CF 10⁻³). MoS, LCR (Cr/Ni/As/Cd/Pb), HQ, HI, and worst-case 95th-percentile SED per Hajrić et al. 2022.

Receptor and exposure-pathway context

The receptor is a 5-year-old child (BW 18.6 kg) applying eye-shadow, nail-polish, or lipstick-type toy makeup to face and eye contour. The SCCS dermal-exposure parameterization assumes 100% bioaccessibility for dermal absorption, application twice daily, and full retention (leave-on). Hand-to-mouth, hand-to-eye, and pica-mediated ingestion are noted in the discussion as additional exposure pathways for toy makeup but are not quantitatively modelled. The thinness of facial skin (~5 mm cheek, ~2 mm eyelid) is cited as elevating absorption compared to other dermal-application sites.

Geographic and seasonal variance

The sample set is geographically restricted to Istanbul toy retailers between an unspecified collection date and the November 2023 submission. 12 of 63 products are Chinese-import (2 of 10 brands); the remaining 51 products from 8 brands are of Turkish origin. The paper does not stratify exceedance ratios by country of origin and does not present seasonal sampling.

Verification notes

  • Cite-key follows the pattern firstauthor-year-location-key-element-count established by similar children’s-product source pages in the wiki.
  • Source uses a closed BRAND A–J anonymization scheme. Per the brand firewall (CLAUDE.md Part 12, strict reading locked 2026-05-17), per-brand-letter ranking and brand-by-brand exceedance lists from Table 1 and Table 4 are deliberately not reproduced; instead, distributions are summarized as across-brand mean ranges, single-sample min–max, and counts of exceedances against named regulatory thresholds. The BRAND-A-4-coded sample with the 171.200 µg/g Pb maximum is reported as “single-sample maximum 171.20 µg/g, 85-fold above the BVL permissible limit” without the brand-letter attribution.
  • The 54.55% / 9.09% Al/Pb SED-vs-TDI exceedance ratios on p. 33985 use a denominator that does not reconcile cleanly with either the 10-brand or 63-sample base (54.55% = 6/11 algebraically; 9.09% = 1/11 or 6/66). The source text reads “n=6” and “n=8” for these two ratios. The Key-numbers entry preserves the source-reported percentages and counts verbatim; downstream synthesis should treat the Al/Pb SED-vs-TDI brand-level exceedance as “majority of brand-mean entries” (Al) and “a minority of brand-mean entries” (Pb) rather than recomputing.
  • The skin-sensitization-limit count for Cr (n=17) on p. 33981 and the LCR-Cd denominator (n=53 for Cd > LOQ) on p. 33979 reflect text-versus-table inconsistencies the source itself does not resolve. The narrative preserves the source’s printed values.
  • The single qualitative organic finding (2-phenoxyethanol) is recorded as detected/not-quantitated per the GC–MS scan limitations stated by the authors. The page does not extrapolate a per-product 2-phenoxyethanol concentration.
  • This source contributes to [[products/childrens-makeup]] (HMTc Cat 2 Row 13) as direct evidence on the cosmetic-pigment platform metals Pb, Ni, Cr (the row’s platform_metals) as well as Al, Mn, Co, Cd, Sb, As, Hg, Zn, Cu, Se, Ba.
  • Audit subagent (2026-05-18) flagged five across-brand-mean-range transpositions in the Key-numbers Table 1 summary where the first-pass row had picked up a brand’s SD or a non-highest brand’s mean instead of the across-brand min/max of the means; corrected against re-read of Table 1 (PDF pp. 33980–33981) to: Al 1121 – 25,117 (BRAND J lowest, E highest); Cr 0.30 – 7.44 (D, C); Mn 0.45 – 151.24 (D, A); Cu <LOQ – 67.30 (D, A); Ba 0.89 – 294 (D, I); As <LOQ – 13.03 (D, A); Cd <LOQ – 0.33 (most brands <LOQ, A 0.33); Pb <LOQ – 81.71 (D, A); Hg <LOQ – 1.32 (most brands <LOQ, I 1.32). The audit subagent’s proposed Ba highest of 91.48 was itself a misread (91.48 is BRAND A’s Ba median, not a mean); independent re-read of Table 1 confirms BRAND I’s Ba mean of 294 µg/g as the actual across-brand maximum.
  • Audit subagent (2026-05-18) flagged the Ni body-table row label as conflating skin-sensitization with permissible-cosmetic-limit exceedance; corrected. The 49/77.78% figure is Ni’s permissible-cosmetic-limit exceedance per p. 33981 (“Cr, Co, and Ni at ratios of 73.02% (n=46), 55.56% (n=35), and 77.78% (n=49)”), making Ni the most prevalent limit-exceeder of the eight regulated toxic elements. The skin-sensitization-limit (>5 mg/kg, Basketter et al.) exceedance for Ni is the separate n=6 / 9.52% figure from p. 33981 and Conclusion. Both are now reported under the correct labels.
  • Audit subagent (2026-05-18) flagged the “7 of 10 brands exceeded Al TDI” attribution to Table 3 worst-case 95th-percentile section as misattributed to Table 3 instead of Table 2 mean SED. Independent re-verification against p. 33985 (paragraph 1) shows the source explicitly attributes this to the worst-case exposure scenario: “According to the worst-case exposure scenario results, 7 out of 10 brands (BRAND A, C, E, F, G, H, and I) exceeded the TDI levels for Al. Also, the worst-case exposure scenario results exceeded the TDI levels in BRAND A for both Pb and As as the prominent toxic elements.” Finding rejected as a false positive — the wiki’s attribution to Table 3 worst-case 95th-percentile is correct per the source text.

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.

CommitDateDescription
ce3e07c2026-05-28activation | Vercel DATACITE env slots set, curators.md filled with founder entry + six scoped reviewer invitations, peer-review onboarding playbook drafted
51400b92026-05-28audit-queue: gasparik2017-wild-boar-slovakia-metals audited-revised