Torres et al. 2024 - Cd, Mg, Pb, Zn in Brazilian commercial soaps and powder detergent

Torres, Nunes, de Paula and Paim developed and validated a wet-decomposition ICP OES procedure for the determination of cadmium, magnesium, lead and zinc in commercial soaps and powder detergent, and applied it to nine products bought in the metropolitan region of Recife (Pernambuco, Brazil). The optimised digestion used 1 mol/L HNO3 plus 30 percent H2O2 with conventional heating on a hot plate at 80 degrees Celsius for one hour; this was compared against muffle-furnace dry decomposition, microwave-assisted wet decomposition, and dissolution in nitric acid followed by direct analysis. Cadmium and lead were below the limit of detection in every sample, while magnesium ranged from 50.9 to 1119 ug/g and zinc from 280.3 to 537.5 ug/g of finished product.

Key numbers

All concentrations are reported on a finished-product mass basis in ug/g (equivalent to mg/kg). The source quantified four elements (Cd, Mg, Pb, Zn) by inductively coupled plasma optical emission spectrometry after wet decomposition with 1 mol/L HNO3 and 30 percent H2O2.

Analytical performance (source Table 2)

Per-sample concentrations (source Table 6)

The selected analytical method yielded Cd < 0.015 ug/g and Pb < 0.5 ug/g across every sample. Magnesium spans more than a 20-fold range across the panel, with the highest level in J6 (the body-moisturizing liquid soap) consistent with magnesium-ion content used dermatologically for skin-barrier function. Zinc clusters between roughly 280 and 540 ug/g, with the highest level in J4 (a bar soap marketed for mixed skin with imperfections).

Cross-method comparison (source Table 5)

For four representative samples (J1, J3, J4, J6) the proposed conventional-heating wet decomposition was compared against microwave-assisted wet decomposition (Milestone HPR-CH-12 reference).

Paired Student’s t-test (95 percent confidence) gave t-values of 0.85 (Mg) and 0.37 (Zn), both below the theoretical 3.18, so the proposed and reference methods were not significantly different for the analytes that were detected. Cd and Pb were below LOD in every sample under both methods, so no paired test was performed.

Recovery (source Table 4)

Spike recoveries were determined at two concentrations per element across four sample matrices (J1, J3, J4, J6). Across all tabulated points, recoveries ranged from 82.3 to 112 percent (mean ± SD shown in Table 4). The abstract summarises recovery as 83 to 119 percent and the Conclusions section states 79.0 to 118.6 percent; see verification notes for the paper-internal discrepancy.

Regulatory comparators cited by the authors

• Brazil ANVISA Resolution RDC No. 44/2012: limits for minerals in artificial organic dyes used in soap manufacturing; in particular, lead at 20 mg/L, other potentially toxic metals up to 100 mg/L, arsenic (expressed as As2O3) 3 mg/L, soluble barium (as barium chloride) 500 mg/L.
• INCI Regulation (EC) No. 1223/2009: zinc pyrithione preservative restricted to a maximum 0.5 percent (5 mg/L) in products intended for brief contact with skin or hair.
• European cosmetic-products Annex VI (last update 2021): maximum 25 percent zinc oxide as a UV filter (nano and non-nano forms).

The authors concluded that the measured values were within the limits established by Brazilian and international legislation.

Methods (brief)

The authors evaluated four sample-preparation strategies before selecting the routine procedure: dry decomposition in a muffle furnace (EDG 3P-S-3000) with a heating step of 30 min at 90 degrees Celsius followed by 60 min at 500 degrees Celsius, with subsequent dissolution of the ash in 1, 2 or 5 mol/L HNO3 with or without 30 percent H2O2; wet decomposition by conventional heating on an open-system hot plate (Fisatom) at 80 degrees Celsius for 1 h; microwave-assisted wet decomposition (Milestone Start D, modified Milestone HPR-CH-12 program, 500 W ramp to 200 degrees Celsius, 45 bar, 15 min); and direct analysis after dilution in nitric acid with sonication (Elmasonic Easy 20 H ultrasonic bath). Each preparation used approximately 0.5 g of sample (3 g for the dry-decomposition variant) and was filled to 25 mL with ultrapure water.

Decomposition quality was assessed by residual acidity (acid-base titration with 0.102 mol/L NaOH and 1 percent phenolphthalein indicator, after Guida et al.) and dissolved organic carbon by ICP OES at the carbon 193.030 nm emission line. The conventional-heating route with 1 mol/L HNO3 plus 2 mL of 30 percent H2O2 was chosen because it gave the lowest residual acidity (0.95 mol/L) and dissolved organic carbon (< 189 mg/L), and produced visually clean digests free of fat layer and particulate residue.

Particle-size distribution of one problem liquid-soap sample (J2) was characterised by dynamic light scattering on a Zetasizer Nano ZS90 (model ZEN3690, Worcestershire, United Kingdom). In 1 mol/L HNO3 the principal peak (93.9 percent of intensity) averaged 524.9 nm with a 6.1 percent tail above 5000 nm; in 2 mol/L HNO3 the principal peak (90.8 percent) averaged 544.8 nm with a 9.2 percent tail at 5381 nm; in 5 mol/L HNO3 a single peak averaged 454.6 nm.

Elemental quantification used a PerkinElmer Optima 7000 DV ICP OES with CCD detector, Echelle optics, concentric nebulizer and Scott-type nebulization chamber, cross-flow nebulizer, axial plasma view, RF power 1.3 kW, nebulizer gas flow 0.8 L/min, auxiliary gas flow 0.2 L/min, plasma gas flow 15 L/min, sample aspiration 1.0 mL/min, and three replicates per measurement. Wavelengths: C (I) 193.030, Cd (I) 228.802, Mg (I) 285.213, Pb (II) 220.353 and Zn (II) 206.200 nm. Argon at 99.999 percent purity was used as the plasma gas.

Calibration standards were prepared from 1000 mg/L stock solutions (Cd, Mg, Pb from Inorganic Ventures; Zn from Fluka Analytical) with curves spanning 0.01 to 8.00 mg/L for the four metals and 10 to 4000 mg/L for carbon (anhydrous D-glucose from Synth). Reagents included HNO3 68-70 percent (Alphatec) and H2O2 (Anidrol). Sample storage and digestion-ware were soaked in 10 percent (v/v) HNO3 for at least 24 h, rinsed with distilled water, and air dried. No certified reference material was available for soap matrices; method accuracy was therefore assessed by spike-recovery at 0.1 and 0.5 mg/L for Cd and Pb and at 1.0 and 6.0 mg/L for Mg and Zn, with spikes added before digestion. LOD and LOQ were calculated according to AOAC guidance.

Two-way analysis of variance (p < 0.05) on residual acidity and dissolved organic carbon, and a paired Student’s t-test (95 percent confidence) on method-to-method element concentrations, were performed in OriginPro 9.0 (OriginLab, Northampton, USA).

The paper does not report Cr, As, Ni, Co, Cu, Fe, Mn, Sb, Sn, Hg or Al concentrations. Chromium speciation, mercury speciation and arsenic speciation are not addressed because those elements were not measured.

Implications

Certification: This source contributes a non-detection record for Cd and Pb in a small Brazilian panel of finished bar soaps, liquid soaps and one powder detergent at LODs of 0.015 and 0.5 ug/g respectively, plus quantitative occurrence values for Zn (and Mg) in the same finished products. The Pb LOD here (0.5 ug/g equivalent to 500 ug/kg or 0.5 mg/kg) is high relative to the more sensitive AAS-based studies in the same product class, so the Pb non-detection should be read as “below 0.5 mg/kg in finished product” rather than as a tight upper bound.

Courses: Useful as a sample-preparation case study for personal-care product matrices, illustrating the trade-off between dry decomposition, hot-plate wet decomposition, microwave-assisted wet decomposition and direct analysis, and how dissolved organic carbon and residual acidity can be used to choose between digestion routes.

App: Contributes Brazil-market context for personal-care soaps and one powder laundry detergent. The panel is small (n = 9, one metropolitan market) and the Pb/Cd LODs are not low enough to characterise typical-product distributions, so the app should treat this source as a Brazil-coverage anchor rather than a distributional reference.

Microbiome: No microbiome data.

Wiki pages this source may touch

• bathing-soaps
• laundry-detergents
• lead
• cadmium
• zinc
• magnesium
• anvisa-rdc-44-2012
• ec-cosmetics-regulation-1223-2009

Verification notes

• Bibliographic metadata verified from the article header on page 1 and the journal masthead: DOI 10.21577/0103-5053.20230164, J. Braz. Chem. Soc. 35(4), e-20230164, 1-10, copyright 2024 Sociedade Brasileira de Quimica, submitted March 31, 2023, published online October 11, 2023, authors Torres, Nunes, de Paula and Paim (Departamento de Quimica Fundamental, Universidade Federal de Pernambuco, Recife-PE, Brazil), corresponding author ana.paim@ufpe.br. The footer of page 1 carries the Creative Commons Attribution License notice.
• Paper-internal discrepancy on recovery range. The abstract states recoveries of 83 to 119 percent; the Conclusions section states 79.0 to 118.6 percent; Table 4 values range from 82.3 to 112 percent (with the maximum recovery being Mg 1.0 mg/L in J6 at 112 percent and the minimum being Zn 1.0 mg/L in J1 at 83.0 percent). The wiki uses Table 4 as the authoritative tabulated range and notes the abstract/conclusions text as an internal mismatch; not silently corrected.
• The abstract states “average hydrodynamic diameter of 525 nm” while the body text gives the principal-peak average in 1 mol/L HNO3 as 524.9 nm (which is the figure quoted here). The abstract’s value is the same to three significant figures.
• Authors do not name brand identities for the nine samples. Sample J labels are retained because they are the source’s anonymous identifiers, not brand names. Product-form descriptors (face, mixed-skin, body moisturizer, antibacterial, bar, liquid, powder detergent) are taken directly from the source’s own descriptive sentences and carry no brand attribution.
• The paper reports magnesium and zinc as quantitative concentrations and cadmium and lead as below-LOD. Magnesium is not a Part 14 priority heavy metal but is retained in the frontmatter metals: list because a magnesium page exists in the wiki and the per-sample magnesium values are needed to interpret why J6 (a body moisturiser) sits at a 20-fold higher level than the other samples.
• Provisional product scaffolds: [[products/bathing-soaps]] is currently a provisional scaffold per the 2026-06-02 ingest of oviri2024-nigeria-bathing-soaps-metals. [[products/laundry-detergents]] is the existing locked product page. Neither is modified by this ingest; routing is the responsibility of the routing audit.
• Missing-slug proposals (Check 2 audit response, 2026-06-03): [[regulations/anvisa-rdc-44-2012]] (Brazil ANVISA Resolução RDC No. 44/2012, Regulamento Técnico Mercosul on permitted colourant substances for personal-care products, cosmetics and perfumes) and [[regulations/ec-cosmetics-regulation-1223-2009]] (Regulation (EC) No. 1223/2009 of the European Parliament and of the Council on cosmetic products) are both cited as regulatory comparators in this paper but do not yet exist as wiki regulation pages. Flagging here as candidates for the regulations directory; not creating the pages in this ingest cycle.
• Matrices vocabulary note (Check 2 audit response, 2026-06-03): the matrices values cosmetic-personal-care, laundry-detergent and household-cleaning-product extend the food-focused common-matrices list documented in docs/gpt-collaboration/system-prompt.md. The same three terms are used in abulude2007-soaps-detergents-akure-nigeria for the analogous Nigerian soap-and-detergent panel, so this is established usage in the source corpus rather than novel invention; flagged here for completeness.
• Evidence Fitness: EF-2 / methodologically rigorous small-panel occurrence study. Limitations: n=9, single metropolitan market, Pb LOD (0.5 ug/g) too high to characterise low-level Pb occurrence, Cd LOD (0.015 ug/g) competitive with comparable studies, and only four metals analysed (Cd, Mg, Pb, Zn) so the page contributes no As, Ni, Cr, Co, Cu, Fe, Mn, Hg, Sb, Sn or Al data for these product classes.

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.