Technical methodology

🇧🇷 Português

Glaze toxicity assessment

This page describes the criteria Keramoslab uses to classify each AI-generated glaze recipe on a scale of 1 to 4. The system combines 3 axes in a decision tree, anchored in the applicable regulatory standards (RDC 42/2013 — MERCOSUR, ISO 6486-2, FDA CPG 7117.07, ASTM C738) and the IARC (WHO) carcinogen classification.

It is a technical screening — guidance, not a verdict.

The score evaluates a formula (the recipe), not the finished product. It exists to give the potter confidence and a documented starting point — the final word on food use is the migration test (ASTM C738) on the actual piece. That test is not legally required: Brazilian norms (like most countries') only set migration limits for lead and cadmium. The other colourant oxides (copper, cobalt, iron, lithium, etc.) have no regulatory limit — several are micronutrients and, in practice, migrate below the body's daily requirement. That is why the “moderate” band does not mean “toxic”: it means “confirm with a test if it will contact food.”

How it works — the 3-axis system

  1. Axis 1 — Oxide hazard: the percentage of each colourant/functional oxide compared to three technical bands (safe / moderate / high) drawn from the international ceramics literature. Prohibited oxides (Pb, Cd, Be) have zero tolerance.
  2. Axis 2 — Chemical stability (Seger UMF): the unity molecular formula is computed for the recipe. The system evaluates the R₂O/RO ratio (fluxes), the glass formers (SiO₂+B₂O₃) as a function of temperature, and alumina (Al₂O₃). A well-vitrified, stable matrix locks colourants in and leaches very little.
  3. Axis 3 — Physical factors: the real finish (glossy/matte) is inferred from the SiO₂:Al₂O₃ ratio on the Stull Chart, not from the user-declared field. Firing maturity is checked against cone/clay-body/atmosphere compatibility.

Score scale (1 to 4)

1/4Food safe

All oxides in the SAFE band, OR up to 3 oxides in the MODERATE band — which, locked into a stable vitreous matrix (Axis 2), are immobilized and protected from leaching. No prohibited oxide (Pb/Cd/Be) and none in the RISK band. Axis 2 (Seger) stable (R₂O/RO ≤ 0.45 with adequate Al₂O₃). A theoretical screening as safe for food contact — confirm with a migration test (ASTM C738) on the real piece. Firing maturity (Axis 3) is a bench-test caveat, not a blocker.

2/4Viable for outer/non-food surfaces

Four or more oxides in the MODERATE band (accumulated load) OR Axis 2 (Seger) borderline/unstable. May be used on the outer part of a utilitarian piece (foot, exterior); for food contact, run the migration test.

3/4Decorative pieces only

An oxide in the RISK (high) band on Axis 1, OR R₂O/RO > 0.45, OR insufficient SiO₂+B₂O₃ for the temperature, OR Stull indicates a matte/unstable matrix. Do not use on pieces that touch food.

4/4Unsuitable for all uses

PbO > 0 OR CdO > 0 OR BeO > 0 (any amount — the only metals with a migration limit in the norms) OR Cr₂O₃ in the risk band under an oxidizing atmosphere (Cr⁶⁺ likely). Reformulate before testing.

Axis 1 — technical bands per oxide

Percentages over the base glaze. References: international ceramics literature + ASTM C738 (leach test) + ANVISA guidelines.

OxideSafeModerateHighTechnical note

Cobalt

CoO

≤1.2%>1.2%–2%>2%Optimal up to ~1.2%; 2% is the ceiling traditionally accepted by potters. Above 2%, excess cobalt tends to leach in acidic media. (Species commonly used in ceramics: CoO.)

Copper

CuO

≤2%>2%–4%>4%Copper is aggressive to the glaze; above 4% leaching becomes perceptible. Copper is an essential micronutrient and is not capped by food-contact migration norms.

Chromium

Cr2O3

≤1%>1%–2%>2%Above 2% it may form soluble chromates (Cr⁶⁺) in acids. IN REDUCTION the limits are more permissive (safe ≤2%, moderate >2%–4%, high >4%) because Cr⁶⁺ does not form.

Manganese

MnO2

≤4%>4%–7%>7%Hazardous by inhalation as raw powder; for food contact prefer ≤4%.

Iron

Fe2O3

≤8%>8%–15%>15%Iron is relatively stable, but excess destabilizes the glaze.

Nickel

NiO

<1%1%–2%>2%Allergen / IARC-listed. Use a stable, well-vitrified matrix; above 2% is flagged for decorative use.

Vanadium

V2O5

≤3%>3%–7%>7%Toxic at high concentrations; stability depends on the glassy matrix.

Titanium / Rutile

TiO2

≤6%>6%–10%>10%Rutile is stable, but excess can create a rough surface that traps soil.

Barium

BaO

≤3%>3%–5%>5%Locked into a stable matrix it is inert; in a weak matrix it leaches. For functional ware, strontium (SrO) is a safer substitute.

Strontium

SrO

≤3%>3%–5%>5%A safer alternative to barium for matte/satin fluxing.

Zinc

ZnO

≤8%>8%–15%>15%Low toxicity — the real risk of excess is a physical defect, caught by the Stull analysis.

Zirconium

ZrO2

≤10%>10%–15%>15%Economical, stable opacifier.

Lithium

Li2O

≤3%>3%–6%>6%Powerful alkaline flux; high amounts lower the matrix's chemical resistance.

Prohibited oxides (zero tolerance)

Lead (PbO), Cadmium (CdO) in any detectable amount — these are the only metals with a food-contact migration limit in the norms (RDC 42/2013, ANVISA Ordinance 27/1996, FDA). Beryllium (BeO) is never acceptable (IARC Group 1). Any presence sends the score straight to 4.

Axis 2 — chemical stability (Seger UMF)

A stable matrix is the real safeguard against leaching: the harder and more vitrified the glass, the less any metal migrates. The system reads three signals — R₂O/RO ratio(excellent ≤0.20, ideal 0.20–0.35, attention 0.35–0.45, high >0.45), glass formers SiO₂+B₂O₃ (adequate range depends on the firing temperature), and alumina Al₂O₃ (ideal 0.25–0.45 mol for high temperature). A commercial frit with no declared composition is treated as stable (it is industrially insoluble by design), not as a failure.

Axis 3 — physical & process factors

The finish is inferred from the SiO₂:Al₂O₃ ratio on the Stull Chart (a glossy, vitreous surface in the ~5–12 zone is best for food). Firing maturity is a bench-test caveat and does not lower the score on its own — cone inference carries ±30–50 °C of uncertainty, so a digital pyrometer reading without a pyrometric cone is never treated as a verdict.

Important — please read before use

  • Referential screening. The score is technical guidance based on chemistry and the literature. It does not replace a leach test performed by a certified laboratory.
  • Every raw-material batch differs. Even following the recipe exactly, variable impurities can affect real leaching.
  • Keramoslab is a guidance platform. It does not sell finished glazes and assumes no legal liability for the result of recipes taken to production.

Questions or technical feedback? Write to suporte@keramoslab.com.br.

🇧🇷 Ler em português