Inductively coupled plasma-mass spectrometry (ICP-MS) detection of trace metal contents of children cosmetics

Overview

RETRACTED ARTICLE. The published PDF supplied for ingest carries a “RETRACTED ARTICLE” watermark on every page. The retraction notice text itself is not part of the supplied PDF, and the retraction date and stated grounds are not visible in this file. This page is preserved as a record of the published claims but evidence tier is downgraded to C and the values must not be pooled into HMTc threshold-setting synthesis or into the consumer-app scoring layer. Any downstream use requires Karen’s review of the retraction notice first.

This open-access Optical and Quantum Electronics study (received 24 Nov 2023, accepted 19 Dec 2023, published online 22 Jan 2024) tested thirty children’s-cosmetics products purchased from the Turkish market for thirteen elemental impurities (Al, chromium (Cr), Mn, Fe, Co, nickel (Ni), Cu, Zn, arsenic (As), Se, mercury (Hg), cadmium (Cd), Pb) by inductively coupled plasma mass spectrometry following microwave acid digestion (HNO₃ 65 % + H₂O₂ 35 %, 9:1). The panel covered 22 eye shadows, 4 lipsticks, and 4 nail polishes drawn from three anonymized “children’s toy makeup kits” (Brand 1, Brand 2, Brand 3) marketed for children in Turkey. The authors compared their findings against US FDA cosmetic limits (Pb 20 µg g⁻¹ for color additives), Health Canada Product Safety Laboratory limits (3 µg g⁻¹ for As/Cd/Hg; 10 µg g⁻¹ for Pb), and the German BfR cosmetic-impurity guidance (Cd 5, As 5, Hg 1, lead (Pb) 20 µg g⁻¹), reporting that “the results of As, Cd, Hg, Pb … are below the limits set by the Food and Drug Administration (FDA) and Health Canada Product Safety Laboratory (PSL) for the 3 brands” — a conclusion that the per-product Table 4/5 values do not uniformly support (see Implications). Evidence tier is C: the article is retracted, the n=1 sample structure per (brand × color) combination limits standalone weight on any single value, calibration range (0-50 µg L⁻¹) is narrow relative to several reported in-product concentrations (notably Al), and the paper presents per-color findings without aggregating to product-form pooled summaries. Retracted-article status overrides any consideration of B-tier on the basis of method clarity.

Key numbers

Thirteen metals quantified by ICP-MS (Agilent 7900), reported as mean ± SD in µg g⁻¹ (= mg kg⁻¹ = ppm) in the as-sold finished product. N=3 replicate analyses per product (analytical replication, not sample replication). Calibration standards prepared at 0-50 µg L⁻¹ in HNO₃ from Merck multi-element stock at 1000 µg L⁻¹. Internal standard solution containing ¹⁴⁵Sc, ¹⁰³Rh, ²⁰⁹Bi at 200 µg L⁻¹ used for sampling-rate and plasma-condition correction. Brand identities are anonymized in the source as Brand 1, Brand 2, Brand 3; the wiki preserves the source’s own numbering without imputing commercial-brand attribution.

Method validation (Table 2 of source, NCS ZC 81002b Human Hair certified reference material)

NCS ZC 81002b Human Hair was analyzed as a CRM to assess digestion and analytical recovery for thirteen elements. Recoveries ranged 94 % (Hg) to 107 % (Cd); all reported recoveries fell within the 94-107 % band.

Method LOD / LOQ (Table 3 of source, µg L⁻¹, N=3)

Eye shadow concentrations across three brands, 22 products (µg g⁻¹, mean ± SD, N=3)

Tabulated from Tables 4 and 5 of the source. Brand 1 carries 10 eye shadows (4 colorful + 6 silvery); Brand 2 carries 5; Brand 3 carries 7.

Lipstick concentrations, 4 products (µg g⁻¹, mean ± SD, N=3)

Nail polish concentrations, 4 products (µg g⁻¹, mean ± SD, N=3)

Note: Table 5 in the source labels the Brand 3 final-row entry “Lipstick / Pink” (the Brand 3 pink lipstick row above) and then continues with what the text describes as four nail polishes. The Brand 3 nail polish (purple) is reported in the text’s panel-maximum summary as part of the four-nail-polish set, with values implicit. The conclusions state “4 nail polish samples” but Table 5 explicitly shows three nail polish rows (Brand 1 pink, Brand 2 pink, Brand 2 purple). The Brand 3 purple nail polish row in Table 5 is the bottom-most row that begins with the “Nail polish” subheading and shows aluminum (Al) 8.2 ± 0.8 — but that same numeric row pattern also appears as Brand 2 purple. Paper-internal inconsistency: the conclusions enumerate “4 nail polishes” but Table 5 as printed shows three nail-polish-form rows for the Brand 2/3 panel. The cell-by-cell row labels in the published table are difficult to verify against the conclusions text. Flagged in Verification notes; values transcribed as printed without imputing the Brand 3 purple nail polish row.

Panel-level maxima as stated in the source’s Results discussion (p. 13)

The source’s results-discussion paragraph (p. 13) states panel-level maxima across the 30 products that do not consistently match the Tables 4-5 cell values. The transcription below cites the table-verified maxima (with the product cell from Tables 4-5 that supports them), not the discussion-paragraph values where those diverge from the tables; paper-internal discussion-vs-table contradictions are documented in Verification notes.

Table-verified panel maxima across all 30 products: Al 822.8 µg g⁻¹ (Brand 1 orange eye shadow), Cr 51.8 (Brand 1 silvery orange eye shadow), Mn 19.2 (Brand 2 pink eye shadow), Fe 67.8 (Brand 1 pink lipstick), Co 0.58 (Brand 1 red lipstick), Ni 1.11 (Brand 2 pink nail polish; Brand 1 pink lipstick Ni 1.1 ± 0.6 is comparable), Cu 9.44 (Brand 3 green eye shadow), Zn 13.3 (Brand 3 orange eye shadow), total arsenic (tAs) 0.84 (Brand 1 purple eye shadow), Se 0.38 (Brand 3 orange eye shadow), Cd 0.26 (Brand 2 pink nail polish), total mercury (tHg) 0.90 (Brand 1 orange eye shadow), Pb 12.2 (Brand 1 green eye shadow).

The authors describe lead exposure across all 22 eye shadows in 3 brands and across colors as falling in the range 0.05 to 12.2 µg g⁻¹ (wet weight basis stated in the discussion text). The maximum Pb 12.2 µg g⁻¹ is below the US FDA 20 µg g⁻¹ cosmetic-color-additive limit (US FDA 2002) but exceeds the Health Canada Product Safety Laboratory 10 µg g⁻¹ Pb limit for cosmetics in three products (Brand 1 green eye shadow 12.2; Brand 1 red lipstick 10.7; Brand 1 pink nail polish 12.10). The source narrative concludes that “the results … are below the limits set by the Food and Drug Administration (FDA) and Health Canada Product Safety Laboratory (PSL) for the 3 brands,” a statement the per-product table values do not support for Pb against the Health Canada limit. Documented as a paper-internal inconsistency in Verification notes.

Methods

Product selection. Thirty cosmetic products were purchased from local markets in Turkey from three different brands of children’s “toy makeup kits” (kits numbered 1-3, brand names not disclosed). The panel composed of 22 eye shadows of seven colors (pink, blue, green, yellow, orange, purple, brown), 4 lipsticks of three colors (red, pink, plus one not described in the sample-collection paragraph), and 4 nail polishes of three colors (pink, purple, and one not described).

Sample digestion. 0.20 g of each cosmetic sample was weighed into 100 mL PTFE digestion vessels in triplicate. An acid mixture of HNO₃ (65 % w/w) and H₂O₂ (35 % w/w) in 9:1 ratio was added. Samples were microwave-digested in an ETHOS Easy™ Microwave Digestion System (Milestone Advanced Digestion Inc., USA) with a two-step program: 1800 W for 20 min at 210 °C, then 1800 W for 15 min at 210 °C. PTFE vessels were cooled to room temperature after digestion. Digested samples were filtered through Whatman No. 40 filter paper to remove undissolved particles, then diluted to 25 mL with double-distilled water before ICP-MS analysis. Blank solutions were prepared in the same manner.

Instrumentation. ICP-MS (Agilent 7900) with autosampler, Babington nebulizer, Ni cones, peristaltic sample-delivery pump. ICP-MS operating conditions: RF power 1550 W, RF matching 1.80 V, sample depth 8.0 mm, torch-H −0.3 mm, torch-V 0.8 mm, carrier gas 1.20 L/min, nebulizer gas 1 L/min, auxiliary gas 0.9 L/min, plasma gas 15 L/min, nebulizer pump 0.1 rps, spray-chamber temperature 2 °C. High-purity argon (99.99 %) was used as the plasma support gas. Pulse/analog detector factor was adjusted daily before analyses. ICP-MS tuning solution at 10 µg L⁻¹ (Ce, Co, Li, Tl, Y) was used for instrument tuning before each experiment.

Calibration. ICP-MS multi-element standard solution from Merck (Darmstadt, Germany). Stock solution and standards diluted in HNO₃. Calibration curve range 0-50 µg L⁻¹. Internal standard solution: ¹⁴⁵Sc, ¹⁰³Rh, ²⁰⁹Bi at 200 µg L⁻¹ for sampling-rate and plasma-condition correction.

Reagents. HCl (37 % w/w) and HNO₃ (65 % w/w) from Merck. High-quality distilled water from a Millipore Milli-Q system.

Reference material. NCS ZC 81002b Human Hair certified reference material (CRM). Same digestion and analysis procedure applied to the CRM to validate the method. Recoveries: 94 % (Hg) to 107 % (Cd). All recoveries within 94-107 %.

Method calibration parameters. LOD = 3 × standard-deviation-of-blank / slope-of-calibration-graph (N=10 repetitive blank analyses); LOQ = 10 × SD / slope.

Speciation. ICP-MS without chromatographic separation. Total arsenic (tAs) and total mercury (tHg) reported; no As-species or Hg-species separation. Total chromium (Cr) reported; no hexavalent chromium (Cr-VI) speciation.

Implications

Certification. The paper would contribute to the Cat 2 Children Personal Care occurrence pool for eye shadows, lipsticks, and nail polishes if the article were not retracted. Per the retracted-article status, the values are excluded from synthesis pools and must not be used in HMTc threshold-setting until Karen has reviewed the retraction notice and confirmed whether any subset of the underlying measurements is salvageable. The page is preserved as a record of the published claims and to make any future re-examination of the values traceable.

Brand-level pattern observation (anonymized). The source’s own discussion observes that Brand 1 carried systematically higher Al, Pb, and Hg concentrations than Brand 2 or Brand 3 across product forms. The three highest Pb values in the panel (12.2 µg g⁻¹ eye shadow, 10.7 µg g⁻¹ red lipstick, 12.10 µg g⁻¹ nail polish) all belong to Brand 1. The three highest Hg values (0.90, 0.65, 0.57 µg g⁻¹) similarly all belong to Brand 1. This per-brand asymmetry is itself a notable finding: in an n=22 eye-shadow / n=4 lipstick / n=4 nail-polish panel, one of three children’s-cosmetic kit sources carries a Pb-and-Hg loading roughly an order of magnitude higher than the other two. If the underlying data are valid, this pattern suggests that the dominant determinant of contamination in this panel is the kit-brand selection (i.e., the manufacturer’s pigment sourcing and quality control) rather than the product form or color. The retraction status forecloses any downstream use until the retraction grounds are known.

Health Canada Pb 10 µg g⁻¹ exceedances. Three of 30 products (Brand 1 green eye shadow 12.2; Brand 1 red lipstick 10.7; Brand 1 pink nail polish 12.10) report Pb above the Health Canada PSL 10 µg g⁻¹ cosmetic-impurity limit. The source’s conclusion that “the results … are below the limits set by … Health Canada Product Safety Laboratory (PSL) for the 3 brands” is inconsistent with the per-product values reported. This inconsistency is documented in Verification notes; it may relate to the retraction grounds.

App. Excluded from app-layer scoring per retracted-article status.

Courses. May be cited in regulatory courses as an example of how retracted analytical papers should be handled in evidence synthesis (preservation as record; exclusion from pooling; cross-reference to retraction notice).

Related evidence

• Lead
• Aluminum
• Mercury, Total
• Cadmium
• Arsenic, Total
• Chromium
• Manganese
• Iron
• Cobalt
• Nickel
• Copper
• Zinc
• metals/selenium
• Eye makeup (liners, shadows, mascara, brow products, false lash adhesives, lash/eyebrow serums)
• Lip products (lipstick, gloss, lip liner, adult lip balm)
• Children's Nail Polish (ages 0-5)
• Children's Makeup (ages 0-5)

Update history

No substantive edit history is available in this build. The full commit record is available in git.