I keep a folder of compliance documents that has crossed my desk more times than any machine catalogue: radioactivity test reports. This article is that folder, opened. I will show you what each number certifies, why China and Europe read the same three measurements differently, and how to spot the certificate that a tender asks for but which cannot legally exist.
TL;DR — Key Takeaways
- Every natural stone carries trace potassium-40, radium-226 and thorium-232; the question is not whether activity exists but which document proves it sits within limits — a GB 6566-2010 Class A report in China, and the Annex XIII index I (Ra-226/300 + Th-232/200 + K-40/3000 ≤ 1) from Directive 2013/59/Euratom in Europe.
- Because the EU divides radium by 300 (China: 370), thorium by 200 (China: 260) and potassium by 3000 (China: 8000), the two rulebooks can reach different verdicts from one gamma spectrum — a K-40-rich granite we screened in 2024 held GB Class A at Iγ = 1.23 yet failed the EU index at I = 1.78.
- A "REACH radiation certificate" does not exist and cannot be issued: REACH is the EU's chemicals regulation, its scope excludes radioactive substances, and Annex V.7 exempts unmodified natural minerals from registration. What EU-bound projects need instead is an accredited gamma-spectrometry report plus an index I declaration.
- A genuine report cannot be produced overnight, because the crushed sample must be sealed for roughly three to four weeks to let radium-226 and its decay products return to equilibrium before HPGe gamma spectrometry can read it.
- Five purchase-order lines — accreditation, method, sampling basis, dual-index calculation and a pre-agreed document list — turn radioactivity compliance from a tender-day scramble into a two-minute file check.
Every piece of natural stone on earth contains traces of three radioactive nuclides — potassium-40, radium-226 and thorium-232 — measured in becquerels per kilogram (Bq/kg). Because the question for a buyer is never whether stone is radioactive but whether a document proves it sits within the limits of the destination market, the two documents that matter are a China GB 6566-2010 test report, which grades material Class A, B or C, and, for Europe, a gamma-spectrum report whose same three Bq/kg numbers feed the activity concentration index I = CRa/300 + CTh/200 + CK/3000 ≤ 1 set out in Council Directive 2013/59/Euratom. What most spec sheets get wrong is the third document: a "REACH radiation certificate," which European tenders still request and which cannot exist, because REACH is a chemicals regulation whose scope excludes radioactive substances and whose Annex V.7 exempts unmodified natural minerals. In this article I open our own compliance folder: how the three nuclides get into stone in the first place, what a GB 6566 report actually certifies, how the EU index reads the identical numbers differently, and the five lines I now write into purchase orders to keep both rulebooks satisfied.
Why Does Natural Stone Contain Radioactivity at All?
I open every screening conversation the same way: radioactivity in stone is not contamination, it is chemistry. Potassium is a major element in feldspar and mica — the minerals that make granite granite — and 0.0117% of all natural potassium is the isotope potassium-40, which is radioactive. Because potassium feldspar is the main reservoir, the more felsic minerals a rock carries, the higher its K-40 reading tends to run — commercial granites commonly sit between 700 and 1500 Bq/kg, while calcite-based marble and limestone, whose crystal lattice has no room for potassium, usually read near zero. Radium-226 (a descendant of the uranium-238 series) and thorium-232 enter the same rocks through accessory minerals — zircon, monazite, apatite — in amounts of tens to a few hundred Bq/kg. The International Atomic Energy Agency classifies all of this as NORM — the same category that covers phosphate rock and coal ash.
Here are the screening ranges from our own compliance folder, accumulated across 187 production batches between 2023 and 2026:
| Stone Family | K-40 (Bq/kg) | Ra-226 (Bq/kg) | Th-232 (Bq/kg) | Typical GB Verdict |
|---|---|---|---|---|
| White/grey marble | 5–40 | 5–20 | 3–12 | Class A, indexes ≈ 0.05–0.15 |
| Beige limestone | 10–60 | 5–25 | 3–10 | Class A, indexes ≈ 0.05–0.15 |
| Travertine | 20–90 | 8–30 | 4–15 | Class A, indexes ≈ 0.05–0.20 |
| Commercial grey granite | 900–1400 | 30–80 | 50–110 | Class A, Iγ ≈ 0.4–0.8 |
| Imported red granite (selected veins) | 1000–1600 | 150–260 | 90–160 | Class A–B boundary; case-by-case |
What a GB 6566 Report Actually Says — and What Class A Buys You
GB 6566-2010, "Limits of radionuclides in building materials," is China's mandatory national standard, in force since July 1, 2011, and every decorative stone sold on the Chinese market — including material we process for export — is graded under it. The report gives you three activity concentrations and two computed indexes: the internal exposure index IRa = CRa-226 / 370, which screens the radium that later decays into radon gas, and the external exposure index Iγ = CRa-226/370 + CTh-232/260 + CK-40/8000, which screens the direct gamma dose from a room fully lined with the material. The class then decides where I may legally sell the stone, and where you may install it:
| Class | Conditions | Permitted Use |
|---|---|---|
| A | IRa ≤ 1.0 and Iγ ≤ 1.3 | Unrestricted — including interiors of homes, hospitals, schools and kindergartens |
| B | IRa ≤ 1.3 and Iγ ≤ 1.9 | Not for interior surfaces of residential, hospital, school or hotel buildings; permitted for industrial interiors and all exteriors |
| C | Iγ ≤ 2.8 | Exterior faces and outdoor uses only |
| Beyond C | Iγ > 2.8 | Not permitted as a decorative building material at all |
Because Class A is defined as safe for the interior of a kindergarten, it is the only class worth specifying for hospitality, healthcare or residential projects — a B-class stone is not "slightly less good," it is legally barred from the exact interiors those projects build. I treat Class A as a non-negotiable entry condition in our own procurement, and so should you.
How the Number Is Produced
I have filed enough legitimate reports to recognize the physical procedure that no express service can shortcut. A sample of at least 1 kg is taken from the body of the slab — at least 300 mm from any edge, per standard sampling practice — then crushed, ground, dried at 105 °C to constant weight and passed through an 80-mesh sieve. The powder is sealed in a calibrated container and left for roughly three to four weeks, because gamma spectrometry reads radium-226 through its short-lived decay products and those must return to equilibrium before the reading is valid. Because the sealed-equilibrium step alone takes about as long as a sea-freight booking window, any supplier who offers a "same-day radioactivity certificate" is offering you fiction. The measurement itself is high-purity germanium (HPGe) gamma spectrometry; I accept only reports carrying CMA (China Metrology Accreditation) and CNAS stamps, with the retained sample kept for at least three years for re-testing.
What Does Europe Actually Regulate? The Euratom Side of the Same Numbers
When I trace Europe's radiological rulebook, it never leads to REACH. It lives in Council Directive 2013/59/Euratom — the Basic Safety Standards Directive — whose Article 75 requires member states to keep the extra annual gamma dose from building materials below a reference level of 1 mSv/year. Because you cannot measure that dose directly at the container port, Annex XIII supplies a screening formula: the activity concentration index I = CRa-226/300 + CTh-232/200 + CK-40/3000, with I ≤ 1.0 marking bulk materials that need no further radiological attention. The index approach was first proposed by the European Commission's Radiation Protection 112 guidance in 1999 and absorbed into law in 2013; I always verify the member-state layer too, because some countries add their own notification or limit requirements.
Two practical consequences follow, and I flag both in every contract review. First, CE marking under the Construction Products Regulation declares mechanical and physical performances — dimensions, flexural strength, water absorption per the relevant harmonized standards — and radioactivity is not part of that Declaration of Performance by default; the radiological file is a separate document chain. Second, because the index assumes bulk use, a thin stone cladding presents less source mass than the formula's worst case, so the index is conservative for our industry — it overprotects rather than underprotects, which is exactly what a screening tool should do.
Why There Is No Such Thing as a "REACH Radiation Certificate"
At least once a quarter a European tender document lands in my inbox asking the supplier to "provide the REACH radiation certificate for the granite." I answer the same way each time: with regret, because that document cannot be issued by anyone, ever — not by an accredited lab, not by a notified body, not by us. The request is a category error with a simple origin: somewhere, a compliance officer merged two European regulations into one file name.
When I walk a client through the mix-up, I start from the fact that the two regimes answer different questions. REACH (Regulation EC 1907/2006) is the EU's chemicals framework — it asks what a substance is made of and whether it is safely managed — and its scope explicitly excludes radioactive substances, which fall under the Euratom Treaty instead. Furthermore, REACH Annex V.7 exempts minerals in their natural state that have not been chemically modified, so a raw marble or granite slab carries no REACH registration duty in the first place. Because REACH governs chemistry and Euratom governs radiation, asking for a "REACH radiation certificate" is like asking the fire brigade to stamp your passport — the officer is real, the stamp is real, but the two systems never meet.
Where REACH does touch our industry is the chemical layer of a stone package: resin-bonded engineered stone, mesh-and-adhesive composite panels, impregnating sealers and installation adhesives are preparations whose substances can carry SVHC notification duties above 0.1% by weight. I provide an SVHC declaration for those components whenever asked — but it says nothing about becquerels, and a gamma-spectrum report says nothing about SVHC.
| Tender Line Asks For | What Exists Instead | Issued By |
|---|---|---|
| "REACH radiation certificate" | Gamma-spectrometry report (Bq/kg) + index I declaration ≤ 1 | CMA/CNAS or ISO/IEC 17025 laboratory + supplier declaration |
| "REACH compliance for resin panels/sealers" | SVHC declaration for the chemical components | Supplier or resin manufacturer |
| "Radiation CE mark" | CE DoP (dimensions, strength, absorption) + separate radiological file | Manufacturer (DoP) + accredited lab (radiological) |
The Case: A Frankfurt Tender, a Missing Certificate, and a Nine-Day Answer
In early 2025 a healthcare developer's tender package for a clinic annex in Frankfurt reached us through the project's stone contractor: 380 m² of grey granite flooring plus 240 m² of marble feature cladding, with the spec requiring, verbatim, a "REACH radioactivity test certificate for all natural stone." The contractor's compliance deadline sat three weeks out. I have learned not to argue with a tender over wording; I have learned to answer the question the drafter meant to ask.
Within nine days the package I sent satisfied the consultant without a single revision request. It contained three sheets. First, the CMA/CNAS-accredited GB 6566-2010 report for the exact granite batch: Ra-226 at 48 Bq/kg, Th-232 at 62 Bq/kg, K-40 at 1050 Bq/kg — IRa 0.13, Iγ 0.50, Class A. Second, a one-page declaration I drafted myself, recomputing the identical numbers into the Euratom Annex XIII index: 48/300 + 62/200 + 1050/3000 = 0.16 + 0.31 + 0.35 = I = 0.82, below the reference value of 1.0. Third, a half-page cover letter mapping each tender clause to the document that answered it, citing the Directive by article and explaining in one sentence why a REACH-issued radiation certificate does not exist. The marble batch needed no drama at all: at I = 0.09 it sat an order of magnitude below every threshold on either continent.
Because the consultant's real worry was a radiological one and the document he named was merely the wrong label for it, the fastest path to acceptance was translation, not argument — and that is the lesson I would hand to any buyer drafting spec sheets: name the regime, name the index, and suppliers can answer in days instead of weeks.
The folder has a failure shelf too, and I keep it deliberately. In 2024 an EU-bound hotel project asked us to consider a dramatic red granite from a specific vein. Our screening read Ra-226 at 210 Bq/kg, Th-232 at 130 Bq/kg, K-40 at 1300 Bq/kg. Under GB it sat at IRa 0.57 and Iγ 1.23 — a legal Class A, just 0.07 under the ceiling. Under the EU index it computed to 0.70 + 0.65 + 0.43 = I = 1.78. I declined to quote it for that destination and steered the design team to a lower-activity alternative, because the same three numbers that earned a green light in one rulebook drew a red flag in the other.
How to Audit a Radioactivity Report in Sixty Seconds
In my experience, fake or worthless reports give themselves away before the numbers do. Here is the checklist I run, in order, every time a factory — ours included — hands me a certificate:
| # | Checkpoint | What Passes | Red Flag |
|---|---|---|---|
| 1 | Accreditation | CMA and CNAS stamps (or ISO/IEC 17025 for non-Chinese labs) | Photocopied stamp, or stamp from a lab you cannot find in the CNAS register |
| 2 | Method line | Names gamma spectrometry and the method standard (e.g., GB/T 11713) | "Tested per international standards" with no method named |
| 3 | Detector | HPGe (high-purity germanium) | No detector named, or a NaI scintillation counter used for certification-grade work |
| 4 | Equilibrium note | Sealed storage of roughly 15–30 days before counting | Report dated two days after sampling date |
| 5 | Units and uncertainty | All values in Bq/kg with ± uncertainty and detection limits | Percentages, "cps," or bare numbers with no uncertainty |
| 6 | Sampling basis | States which quarry, batch or production lot the core came from, ≥ 1 kg crushed core | "Representative sample" with no origin, or sample ID matching a different material |
| 7 | Report date | Within 12 months, or re-issued after any quarry or vein change | Identical values to last year's report to the becquerel — template, not measurement |
Run the Numbers Yourself: A Dual-Standard Translator
Because the two rulebooks use the same inputs and different denominators, you do not need a consultant to pre-check a stone — you need arithmetic. Move the sliders to any report's three values and the translator computes both verdicts side by side. I have preloaded three real batches from our folder as starting points.
Dual-Standard Radioactivity Translator
China · GB 6566-2010
EU · Annex XIII Index
What the Certificate Still Cannot Tell You
I would rather tell you the edges of the map than let you fall off it. A radioactivity report is a screening document, and I mark four things outside its border:
It does not measure radon exhalation. The internal exposure index screens radium-226, the parent of radon-222, but the gas that actually escapes a finished slab depends on porosity, surface finish and sealing — a gamma spectrum reads the parent inventory, not the exit rate. Indoor radon is a room-air problem governed by the WHO's 100 Bq/m³ reference level, and in most buildings the dominant radon source is soil gas through the foundation, not the stone on the walls. Because a Class A certificate answers the material question while radon remains a ventilation question, a healthcare or basement-level project that worries about radon needs an air measurement, not a better stone certificate.
It does not compute your room's actual dose — I trust the index for screening and nothing further. The indexes assume a standard room lined with the material; a 20 mm feature wall presents far less source mass than that worst case. The index is deliberately conservative — but conservative is not the same as exact.
I never treat one certificate as covering every slab. Activity concentrations vary within a quarry and even within a block. Our screening works because we test per batch entering production — 187 batches in three years — not because one archived report covers a material forever. Because a vendor can photograph one good report and ship a different vein, the sampling basis on the document (checkpoint 6) is what ties the paper to your stone.
Finishing chemistry changes the edge cases, not the core. Resin impregnation or a mesh back on a fragile slab adds milligrams-per-square-metre of foreign material — irrelevant to radiation, occasionally relevant to REACH as a chemical component. Polished, sealed surfaces measurably reduce exhalation; they do not change the Bq/kg inside the slab, which is what both indexes grade.
Five Lines Worth Adding to Your Purchase Order
These are the clauses I write when I sit on the buyer's side of the table, and the ones I answer fastest when someone sends them to me. Every material family in the order — from the calcite marbles across our slab range to any granite sourced on request — can be made to satisfy all five without heroics:
1. Recency and accreditation: "Supplier shall provide a GB 6566-2010 radioactivity test report, issued within the last 12 months by a CMA/CNAS-accredited laboratory, for each stone family in this order."
2. Method transparency: "The report shall state the test method (gamma spectrometry), the detector type, the sealed-equilibrium period, and the measurement uncertainty for each nuclide."
3. Sampling basis: "The tested sample shall be identified by quarry and production batch, taken as a ≥ 1 kg crushed core from the slab body, and the retained sample shall be available for re-testing."
4. Dual-index calculation: "For EU-bound material, the supplier shall additionally declare the Annex XIII index I computed from the same measured values; the declared value shall not exceed 1.0."
5. Document mapping: "Where the project specification names a document that does not exist under EU law (e.g., a 'REACH radiation certificate'), the parties agree that the GB 6566 report plus the index I declaration satisfy the radiological requirement, per Directive 2013/59/Euratom Article 75."
Frequently Asked Questions
Send Us the Tender Clause That Confuses You
If your specification sheet asks for a certificate that seems not to exist — or a report you cannot decode — forward it to us. I will map every clause to the document that satisfies it, with real batch data behind each number, before you commit a deposit.
Sources: GB 6566-2010 (Limits of radionuclides in building materials); Council Directive 2013/59/Euratom, Articles 75–76 and Annex XIII; Regulation (EC) No 1907/2006 (REACH), scope and Annex V.7; IAEA NORM programme documentation; WHO indoor radon guidance; Ruifengyuan internal radioactivity screening records, 2023–2026. Written by Alex Wang, International Business Director, Ruifengyuan Stone, Shuitou, Nan'an, Fujian, China.
Post time: Nov-10-2026