Matei et al. 2025 — Cd, Pb, Cu, Cr, Co, Mn, Ni in eight Romanian edible vegetable oils by GF-AAS

This study determined seven trace metals (Cd, Pb, Cu, Cr, Co, Mn, Ni) and seven physicochemical quality indices (acidity index, moisture content, peroxide value, refractive index, iodine value, saponification value, total phenolic compounds) in eight types of edible vegetable oil purchased from local markets in Constanta, Romania, in 2024. Three samples of each oil type were analyzed (24 samples total) by graphite furnace atomic absorption spectrometry after microwave-assisted nitric acid/hydrogen peroxide digestion. Across all eight oils, Cd was 0.001–0.010 mg/kg, Pb 0.011–0.217 mg/kg, Cu 0.010–0.672 mg/kg, Cr <LD–0.062 mg/kg, Co 0.003–0.031 mg/kg, Mn 0.007–0.083 mg/kg, and Ni 0.003–0.015 mg/kg. The authors report that three values exceed the 0.1 mg/kg FAO/WHO 2002 ceiling cited for Cu in vegetable oils (blend oil 0.672, organic extra virgin olive 0.312, sunflower not exceeding) and for Pb (rice oil 0.217, organic extra virgin olive 0.101 at the cap). Target hazard quotient (THQ) and hazard index (HI) values were below 1 for all oils (HI range 0.00369–0.02672), and the sum of incremental lifetime cancer risks (∑ILCR, for Cr, Cd, Pb, Ni) ranged 4.41×10⁻⁶ to 16.39×10⁻⁶ for adults — within the U.S. EPA acceptable range of 1×10⁻⁶ to 1×10⁻⁴. The paper measures total Pb, total Cd, and total Cr (no speciation); arsenic and mercury were not measured.

Note on evidence tier: assigned B. Peer-reviewed MDPI Applied Sciences with a complete methods description, calibration and digestion documentation, but per-oil-type n is small (n=3) and the geographic frame is a single Romanian city (Constanta) market in a single year (2024). The seven-metal panel omits the two HMI-priority analytes most regulated in food-safety contexts (arsenic and mercury). Total Pb, total Cd, and total Cr are reported without speciation, which is appropriate for these elements in oils.

Key numbers

All metal concentrations are reported in mg/kg as mean ± standard deviation, average of three samples per oil type (Table 2, page 6). The basis is the oil as purchased (no wet/dry-weight distinction is made in oils literature). Physicochemical values are reported as mean ± SD in their native units (Table 1, page 6).

Physicochemical quality indices (Table 1):

Metal concentrations (Table 2, mg/kg ± SD; n=3 per oil type):

Across-oil ranges (text and Table 2):

• Ni: 0.003–0.015 mg/kg (least abundant alongside Cd).
• Cd: 0.001–0.010 mg/kg.
• Pb: 0.011–0.217 mg/kg.
• Cu: 0.010–0.672 mg/kg (lowest in extra virgin olive 0.010; highest in blend oil 0.672).
• Mn: 0.007–0.083 mg/kg.
• Cr: <LD (grapeseed)–0.062 mg/kg (lowest in extra virgin olive below LOD; highest in blend oil 0.062).
• Co: 0.003–0.031 mg/kg.

Target hazard quotient (THQ) per metal (Table 3, mg/kg/day; Σ THQ per oil):

Incremental lifetime cancer risk (Table 4, ×10⁻⁶) computed for Cr, Cd, Pb, Ni only (cancer slope factors 0.5, 0.38, 0.085, 0.84 kg·day/mg respectively, ADAF = 1 for adults):

Exposure parameters used in the THQ and ILCR calculations (page 4): ingestion rate IR = 25 g/day; exposure frequency EF = 365 day/year; exposure duration ED = 70 years; body weight BW = 70 kg; averaging time AT = 25,550 days. Oral reference doses RfD (mg/kg/day): Pb 5×10⁻⁴, Cu 4×10⁻², Cd 1×10⁻³, Cr 3×10⁻³, Co 2×10⁻², Mn 2.4×10⁻¹, Ni 2×10⁻². U.S. EPA carcinogenic risk acceptable range cited: 1×10⁻⁶ to 1×10⁻⁴.

Pairwise Pearson correlations (Table 6, p<0.1 threshold; bold-significant correlations from the paper):

• Co–Mn r = 0.97 (strongly positive, significant; authors infer common soil origin).
• Cu–Cr r = 0.72 (significant).
• TPC–Pb r = −0.64 (significant negative); TPC–Ni r = −0.51 (significant negative).
• PV–Mn r = 0.63 (significant positive).
• SV–Ni r = −0.66 (significant negative).
• AI–IV r = −0.65 (significant negative).

Regulatory benchmarks cited by the authors:

• FAO/WHO (2002) maximum permissible heavy-metal content in oils: 0.1 mg/kg for Cu and Pb (page 9, citing reference 59). The authors state there is currently no Codex legislation for Cd in oils (page 9, citing references 57, 58 — FAO/WHO 2002 and EU 2023/915).
• Codex Standards for Edible Oils (CXS 210-1999): acidity index limit 0.3 mg/g for olive and nut oils (page 7); peroxide value 10 meqO₂/kg cap; iodine value 132–162 g I₂/100 g for most edible oils; saponification value 188–196 mgKOH/g for linseed, 184–196 for olive, 168–181 for rapeseed, 189–195 for sunflower (citing references 18, 36, 42).
• WHO recommended moisture content for edible vegetable oil: 0.2% (page 7, citing reference 36).
• Codex refractive index range cited (1.463–1.471) for high-quality oils (page 8, citing reference 39).
• 915 cited generally for cadmium limits in foods (reference 58); the paper does not name a specific vegetable-oil Cd limit.
• For ILCR interpretation the authors cite the U.S. EPA reference dose handbook (reference 25), the EPA Regional Screening Level table (references 23, 27), and EPA carcinogenicity guidance (reference 28).

Exceedances and notable findings (Discussion, sections 4.2 and 4.3):

• Cu exceeded the cited 0.1 mg/kg FAO/WHO ceiling in blend oil (0.672 mg/kg) and organic extra virgin olive oil (0.312 mg/kg).
• Pb exceeded the cited 0.1 mg/kg FAO/WHO ceiling in rice oil (0.217 mg/kg); organic extra virgin olive oil reached the limit at 0.101 mg/kg.
• Walnut oil’s acidity index (2.080 ± 1.649% oleic) is above the 0.3 mg/g Codex cap for olive and nut oils.
• Walnut oil’s iodine value (72.7 ± 5.744 g/100 g) is below the 132–162 g I₂/100 g range cited.
• All oils’ saponification values were below the cooking-oil ranges the authors cite for sunflower, olive, rapeseed, and linseed oils.
• Moisture content was below the WHO 0.2% recommendation for all oils.
• HI < 1 for all oils, indicating non-carcinogenic risk is acceptable. ΣILCR for the four cancer-relevant metals was within the EPA acceptable range (1×10⁻⁶ to 1×10⁻⁴) for all oils.

Methods (brief)

Sampling: 24 samples (eight oil types × three samples each) of sunflower, grapeseed, extra virgin olive, organic extra virgin olive, walnut, rice, rapeseed, and blend (flax + rapeseed + pumpkin) oils purchased from local markets in Constanta, Romania, in 2024 (page 3, section 2.1).

Physicochemical methods (section 2.2):

• Acidity index (AI): 5 g oil dissolved in 1:2 alcohol-ether, titrated with 0.1 N NaOH to phenolphthalein endpoint; expressed as mg of oleic acid (AOAC).
• Moisture content (MC): 5 g sample in pre-tared vial, oven at 105 °C to constant weight; %w/w.
• Peroxide value (PV): 1 g sample in 250 mL conical flask with 10 mL acetic acid/chloroform 3:2, 1 mL saturated KI, 2 mL 1% starch, titrated with 0.1 N sodium thiosulfate; blank in parallel; meqO₂/kg.
• Refractive index (RI): Abbe NAR-1T solid refractometer (Atago, Kobe, Japan), 20–25 °C, per AOAC 1984.
• Iodine value (IV): 1 g sample in CHCl₃ solvent + 25 mL Hanus reagent, dark/60 min, 20 mL 15% KI, titrated with 0.1 N Na₂S₂O₃ in presence of starch; g I₂/100 g.
• Saponification value (SV): 2 g oil + 25 mL 0.5 N alcoholic KOH, reflux 1 h, excess KOH titrated with 0.5 N HCl with phenolphthalein; mgKOH/g.
• Total phenolic compounds (TPC): ~10 g sample macerated 1 week in 100 mL 98% ethanol in the dark; filtered (Whatman); 1 mL sample + 1 mL Folin–Ciocalteu reagent (Merck) + 1 mL 20% sodium carbonate (Redox, Otopeni, Romania); absorbance read at 675 nm against blank on a Jasco 550 UV-VIS spectrometer (Jasco International, Tokyo, Japan); calibration y = 0.0699x, R² = 0.9996; expressed as mg gallic acid equivalent (GAE, Fluka, Buchs, Switzerland) per kg.

Trace metal determination (section 2.2.8): Microwave-assisted digestion using a Berghof Speedwave ENTRY DAP-60K system (Thuringia, Germany). 0.2 g oil + 10 mL ultrapure HNO₃ 65% (Merck KGaA, Darmstadt, Germany) + 2 mL H₂O₂ 30% (Merck KGaA, Germany) in PTFE vessels; 10 min rest before sealing; digestion program 200 °C (5 min ramp to 160 °C; 3 min at 200 °C and held at 160 °C for 5 min; 200 °C for 10 min; 10 min at 75 °C). Digests diluted to 50 mL with deionized water and filtered through 0.45 µm. Quantification on an Analytik Jena ContraA 800D graphite-furnace atomic absorption spectrometer (the paper labels the method “GF-AAS” / “GTAAS”; Jena, Thuringia, Germany). Calibration from a multi-element ICP-standard solution IV (Merck KGaA, Germany); deionized water from a Direct Q UV Millipore system (Burlington, MA, USA; ~18.0 MΩ; Analytik Jena Instruments). Performance parameters of the analytical method are cited from prior published work by the authors (reference 22) and are not reproduced in this paper.

Health risk assessment (section 2.3):

• THQ = EDI/RfD = (C·IR·ED·EF)/(BW·AT·RfD). EDI = estimated daily intake; C = metal concentration in oil (mg/kg); IR = 25 g/day; EF = 365 day/year; ED = 70 years; BW = 70 kg; AT = 25,550 days. RfDs as listed above.
• HI = ΣTHQ across the seven metals per oil.
• ILCR = EDI × CSF × ADAF. CSFs (kg·day/mg): Pb 0.085, Cd 0.38, Cr 0.5, Ni 0.84. ADAF = 1 adults, 3 children. CSFs for Co, Cu, and Mn do not exist; ILCR cannot be computed for those metals.
• ΣILCR = ILCR_Pb + ILCR_Cd + ILCR_Cr + ILCR_Ni. EPA acceptable carcinogenic risk: 1×10⁻⁶ to 1×10⁻⁴.

Statistics: Pairwise Pearson correlations computed in R 4.2.1 with the “cooplot” R package. Significance threshold p < 0.1 (Table 6 footnote).

No speciation: total Pb, total Cd, total Cr (not Cr-VI), total Co, total Mn, total Cu, total Ni. Arsenic and mercury were not measured.

Implications

• Certification (HMTc): contributes Romanian-market occurrence data for Pb, Cd, Cu, Cr, Co, Mn, and Ni in eight edible-vegetable-oil categories (n = 3 samples per category), sampled from Constanta supermarkets in 2024. The two oil categories with multiple metals approaching or exceeding the 0.1 mg/kg FAO/WHO ceiling cited by the authors are the seed-blend oil (Cu 0.672, Pb 0.099, Cr 0.062 mg/kg) and the organic extra virgin olive oil (Cu 0.312, Pb 0.101 mg/kg). Rice oil contributed the highest single-metal value (Pb 0.217 mg/kg) and the highest single-metal THQ (Pb THQ 22.14×10⁻³). At per-oil ∑ILCR up to 16.39×10⁻⁶ (blend oil), the cumulative cancer risk remains within the EPA 1×10⁻⁶–1×10⁻⁴ acceptable band for the four cancer-relevant metals at the modeled 25 g/day adult intake. No As or Hg data are contributed by this study.

• Courses: useful as a Romanian-market companion to Turkish (acar2012-vegetable-oils-olives-turkey-etaas, pehlivan2008-vegetable-oils-turkey-icp-aes), Ethiopian (yohannes2024-vegetable-oils-ethiopia-gondar), Iranian (mehri2024-vegetable-oils-iran-hamadan-ptes), Saudi (ashraf2012-heavy-metals-vegetable-oils-saudi-arabia), and Taiwanese (lin2024-edible-oils-taiwan-12-metals) edible-oil studies. It is also a worked example of (1) full THQ/HI/ILCR risk-assessment arithmetic from raw concentration tables, (2) Berghof Speedwave microwave digestion conditions and the Analytik Jena ContraA 800D GF-AAS platform for oil matrices, and (3) Pearson correlation analysis between physicochemical indices and trace-metal concentrations to suggest oxidation-related or shared-source signal.

• App: oil-side Pb and Cu inputs from this paper are most relevant for ingredient-list-driven exposure estimators that disambiguate edible-oil subtype. Per-serving estimators for rice oil and seed-blend oils should incorporate the upper-range Pb (0.217 mg/kg) and Cu (0.672 mg/kg) values; for organic extra virgin olive, the upper-range Cu (0.312 mg/kg) and Pb (0.101 mg/kg) values. Cd values are uniformly low (≤0.010 mg/kg).

Wiki pages this source may touch

• olive-oil
• sunflower-oil
• rapeseed-oil
• rice-bran-oil
• walnuts
• vegetable-oils
• olive-oil
• cooking-oils-other
• lead
• cadmium
• copper
• chromium
• cobalt
• manganese
• nickel

Verification notes

• 2026-06-01 Fresh ingest from raw/Manual Fetch Kimi /June 1/applsci-15-06269.pdf. Three identity checks (DOI 10.3390/app15116269 grep, raw_handle MFK_applsci-15-06269 grep, cite-key matei2025* grep) all returned no existing pages.
• The paper labels its trace-metal method “Graphite Atomic Absorption Spectrometry” / “GTAAS” and uses an Analytik Jena ContraA 800D instrument. ContraA 800D is a high-resolution continuum-source AAS that supports graphite-furnace operation; the paper uses the GF mode. The methods section captures the paper’s “GF-AAS” / “GTAAS” labels verbatim.
• “Rice oil” is the paper’s term throughout. Rice bran is the only commercial source of edible rice-derived oil, and the paper’s own comparison literature ([Mohajer 2019, ref 61; Bakota et al. 2015, ref 68]) is explicitly about rice bran oil. The page uses rice-bran-oil in the ingredients frontmatter to route the rice-oil signal to the appropriate existing wiki page; matrices uses bare rice-bran-oil for consistency.
• The blend oil is a flax + rapeseed + pumpkin-seed combination. Flaxseed-oil and pumpkin-seed-oil ingredient pages do not exist in the current taxonomy snapshot. Routing the blend-oil signal to rapeseed-oil (the one shared component with an existing wiki page) and to vegetable-oils preserves traceability without overspecifying. Flagged for Karen: a flaxseed-oil (alias linseed-oil) ingredient page may be worth creating once a second source measuring flax-derived oil is encountered.
• “Grapeseed oil” and “walnut oil” do not have dedicated ingredient pages in the taxonomy snapshot. walnuts is used to route walnut-oil contamination data to the existing walnuts ingredient page (the closest taxonomic anchor for the same plant). Grapeseed-oil signal is routed to vegetable-oils only. Flagged for Karen: grapeseed-oil and walnut-oil ingredient pages may be worth creating once a second source measuring these specifically is encountered.
• Matrices uses bare strings (olive-oil, sunflower-oil, rapeseed-oil, walnut-oil, grapeseed-oil, rice-bran-oil, blend-oil, edible-oil) following the convention established in acar2012-vegetable-oils-olives-turkey-etaas and pehlivan2008-vegetable-oils-turkey-icp-aes. Per the audit note on those pages, bare-string matrices outside the system-prompt list are advisory only.
• Authors measure total Pb, total Cd, and total Cr (no speciation). Frontmatter uses Pb, Cd, Cr accordingly (not iAs/tAs/MeHg-style speciation markers; Cr is not Cr-VI).
• Reagent and instrument vendor names (Merck KGaA, Berghof, Analytik Jena, Atago, Jasco, Fluka, Redox, Millipore) are preserved per Part 12 Exception 2 (scientific-method vendor names). No sampled-product brand names appear in the paper; the eight oil categories are labeled by type and provenance only.
• The paper cites several regulatory benchmarks (FAO/WHO 2002 limits at 0.1 mg/kg Cu and Pb; Codex CXS 210-1999 for physicochemical indices; WHO 0.2% moisture cap; EU 2023/915 for Cd in foods generally). Of these, only eu-2023-915-cadmium is in the current taxonomy snapshot; the other limits are preserved in prose without wikilinks rather than asserted as cross-page facts.
• Author affiliation: Department of Chemistry and Chemical Engineering, Ovidius University Constanta, 124 Mamaia Blvd., 900527 Constanta, Romania. Correspondence: sbirghila@univ-ovidius.ro (S.B.); asoceanu@univ-ovidius.ro (A.S.).
• Article metadata: Applied Sciences 2025, 15, 6269. Received 23 April 2025; revised 28 May 2025; accepted 2 June 2025; published 3 June 2025. DOI 10.3390/app15116269. CC BY 4.0 (MDPI standard, confirmed in copyright statement: “open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license”).

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.