You can specify zirconia. You cannot specify your way out of the supply chain that makes it. Zircon sand — the mineral commercial zirconia starts from — is mined as a co-product of titanium feedstock, so your mine can be idled for reasons unrelated to ceramics. And the material crosses a statutory radioactivity threshold in Korea and the EU, or does not, depending on the refining route rather than on the word "zirconia". Three conclusions first: name the refining route, ask which convention the ZrO₂ percentage uses, and get a lot gamma-spectrometry certificate in Bq/g before the material ships.
The mine answers to a different market
The United States Government states the coupling in a table. The Federal Register notice publishing the Final 2025 List of Critical Minerals on 7 November 2025 carries a by-product column reading, verbatim, "Zirconium — Yes — Titanium, tin" and "Hafnium — Yes — Zirconium." USGS says it in prose in Minerals Yearbook 2021: zircon is "a coproduct of the mining and processing of heavy-mineral sands for the titanium minerals ilmenite and rutile."
That coupling fired inside the last twelve months. Iluka Resources' Annual Report 2025 records that, "in response to weaker demand in the titanium dioxide market, especially relating to pigment," production at the Cataby mine and Synthetic Rutile kiln 2 was suspended from 1 December 2025; by the March 2026 quarter the mine was idled, "with no production of zircon or rutile". Cataby sits in Iluka's Perth Basin reserve region, where the heavy-mineral assemblage of the ore reserve runs 59% ilmenite against 12% zircon. A decision taken inside the paint-pigment market removed a zircon stream.
But "zircon is a mere by-product" is only half right. Iluka's Eucla Basin reserve assemblage runs 45% zircon against 26% ilmenite, and on our own conversion of its 2025 accounts — approximate, since the relevant lines are published in different currencies — zircon was roughly three-fifths of mineral-sands revenue on three-tenths of the tonnes. The coupling is strong at mines built for titanium and weak at mines built for zircon; the buyer does not choose which one idles.
Mining is less concentrated than usually quoted: USGS Mineral Commodity Summaries 2026 puts 2025 world production of zirconium mineral concentrates at an estimated 1,200 kt gross weight, of which Australia 400 and South Africa 270 — 670 kt, or 55.8%, and Mozambique 160.
What sets the price is ceramic tile, refractories and foundry sand, not advanced ceramics. Iluka's H1 2026 results — which reached us through trade-press reporting, not a filing we could open — called the market "a supply-side story, with industry supply outages and challenges constraining high-quality zircon availability," while its quarterly commentary names Chinese, European and Indian tile demand as the swing factors. (We put no tonnage on a 2026 deficit, having traced none to a primary source.) A specification buyer of zirconia powder is a price-taker in a market that does not know they exist.
The crystal has to be destroyed, and that step decides the radiology
Zircon is zirconium silicate, ZrSiO₄; about a third of it by mass is silica, and nothing high-purity comes out until the crystal is broken. The two routes, described in J.H. Selby's IAEA NORM paper written from inside the zircon industry, diverge here:
- The chemical route does not dissociate the zircon. Caustic soda fusion at 700 °C gives sodium zirconate, dissolved in hydrochloric acid to form zirconium oxychloride (ZOC), which becomes zirconia by calcination or by hydroxide precipitation.
- The thermal route dissociates zircon into ZrO₂ and SiO₂ above 1800 °C, after which caustic soda dissolves out the silica; the acid variant instead dissolves the zirconium, later precipitated as the oxide. This route, Selby notes, "retains the high boiling-point elements such as uranium and thorium."
ZOC is where the geography narrows, including for buyers who believe they have avoided it. Tosoh's published flow for its TZ-series powders runs zircon sand → ZrOCl₂·8H₂O (+ YCl₃) → hydrolysis → drying → calcination → milling → spray drying → zirconia powder. Buying a Japanese powder changes who does the hydrolysis; it does not remove the oxychloride step, and where a given maker sources its ZOC is a fair question to put to any of them.
How concentrated is that step? The figure in circulation — China at roughly 95% of world ZOC output — traces only to paywalled market-research abstracts, and we do not repeat it as fact. What is primary is a contrast inside one country's customs data: USGS reports US imports of zirconium ores and concentrates over 2021–24 as South Africa 48%, Australia 35%, Senegal 15%, with China absent, while zirconium compounds were China 41%, South Africa 31%, France 12%. China was also the leading global importer of concentrates. Sand flows one way, compound the other.
One trap goes with it. A listed Chinese zirconium-chemicals producer's audited 2025 annual report, citing China Customs, gives 2025 zircon sand imports of about 2.14 million tonnes against zirconium oxychloride exports of about 58,000 t. Do not set that tonnage beside the USGS production figure and conclude someone is wrong: USGS counts concentrates on gross weight, and notes that China's imports include zircon in mixed and separated heavy-mineral concentrates that it does not count as zircon concentrate. Tonnage declared under HS 2615.10 is not tonnage of zircon.
Radioactivity is a procurement obstacle before it is a physics question
Uranium and thorium substitute for zirconium in the zircon lattice. Selby gives most commercial zircons at 250–350 ppm uranium and 100–200 ppm thorium, generally in secular equilibrium — a ²³⁸U activity concentration of 3–4.5 Bq/g. An independent measurement points higher: Hazin and colleagues at Brazil's CRCN-NE, by gamma spectrometry on zircon sand products from a Brazilian plant, report ²³⁸U at 4.5–14.0 Bq/g and ²³²Th at 0.9–7.9 Bq/g. Neither is in error: zircon radioactivity is deposit-specific, and a sand carrying separate grains of monazite does not behave like one without them.
Now the lines those numbers must clear.
Korea's thresholds are numerically identical to the EU's — both derive from the IAEA exemption values shown in the table above; we did not verify a drafting history and assert none. Under 제9조제1항 the first listed registrant is a person who mines, imports or exports, or sells raw material (원료물질) — an importer of zircon sand is inside that description. 시행령 제4조제1항 requires an instrument able to measure activity concentration or dose rate at the handling location, and the Act's 제11조제1항 requires notification to the Nuclear Safety and Security Commission on every import or export.
The quantity threshold is written as a total activity, not a mass. Dividing 1,000 kBq by the activity concentration:
On typical commercial zircon a Korean site reaches the quantity trigger at roughly a quarter to a third of a tonne a year — before the first bulk bag is empty.
Two caveats. The text we read does not settle whether the 1,000 kBq applies per radionuclide or to an aggregate; the table uses ²³⁸U alone, the conservative reading. And it is a statutory reading only — we could not verify how the NSSC treats imported zircon sand in practice, having searched for a register of 원료물질 handlers and for zircon-specific guidance and found neither. Confirm with the Commission or a licensed radiation institution, not on our reading or a supplier's word.
Transport is usually the smaller problem: ordinary zircon at 3–4.5 Bq/g sits under the 10 Bq/g exclusion, but Hazin's high sample sums to about 21.8 Bq/g across ²³⁸U and ²³²Th. The exclusion is deposit-specific, not a property of zircon.
Then the finding that decides your exposure. On Selby's figures, ²³⁸U activity is 3–4.5 Bq/g in raw zircon, below 0.1 Bq/g in chemical-route zirconia, 1–2 Bq/g in process waste streams, and up to 5,000 Bq/g in tank residues. The activity does not vanish in refining; it moves into the effluent. But thermal-route zirconia keeps it — fused zirconia runs 4–6 Bq/g as abrasive feed and refractory products about 2.5 to 5 Bq/g, while a producer starting from purchased ZOC would probably not need to notify, since its raw material is already below 1 Bq/g. So the ladder is:
- Zircon sand or milled flour — you are handling 원료물질; expect the registration and per-shipment notification questions.
- Fused or electro-fused zirconia — still likely above 1 Bq/g.
- Chemical-route high-purity ZrO₂ or YSZ powder — below the line on Selby's figures.
Note the asymmetry: the fused grade is the volume grade — the same Chinese filing puts it at over half the raw material for ceramic colours and glazes — while the chemical grade is the specification grade. So ask for a lot-specific gamma-spectrometry certificate in Bq/g for ²³⁸U, ²³²Th and ⁴⁰K, not a generic statement, in the same routine as our SEM/EDS lot verification.
"99.9% ZrO₂" is a number about two elements
Hafnium sits in the zircon lattice with zirconium and comes out with it. USGS gives the ratio as about 50 to 1 in one publication and 34 to 1 in another, so use the range. On the oxide basis that puts HfO₂ at roughly 1.7–2.5 wt% of commercial zirconia, and is not removed, which is why industry writes purity as "ZrO₂ + HfO₂". Tosoh's published TZ-series specification allows HfO₂ below 5.0 wt% across every grade, and Sigma-Aldrich's entry for zirconium(IV) oxide at 99.99% trace-metals basis states the assay "excludes ~2% HfO₂". A 99.99% material can legitimately contain 2% of another element. Ask which convention the certificate of analysis uses before comparing quotes — the same problem that makes the "ZrO₂ balance" line on a coating powder datasheet include hafnia, covered in our TBC powder specification guide.
Do not ask for less hafnium: removing it needs solvent-extraction chemistry at hafnium economics (unwrought hafnium averaged US$3,800/kg in 2025 against zirconium sponge at US$22/kg) and has a regulatory consequence. US export control ECCN 1C234 covers "zirconium with a hafnium content of less than 1 part hafnium to 500 parts zirconium by weight," and explicitly reaches compounds, not only metal. That threshold is 0.2 wt% of the zirconium; the hafnium in natural zirconium runs 2.0–2.9% of the zirconium by weight, ten to fifteen times the control line, and is not caught. A buyer who writes "hafnium-free" or "Hf < 100 ppm" into a specification without needing it has written themselves into a licence conversation. We verified the US text only, not Korea's 전략물자수출입고시 or China's dual-use list.
We also looked for a published ASTM or ISO standard for zirconia powder purity grades and could not find one; a specification must be built from named limits, not a grade name. The grades we hold specifications against are listed under zirconia ceramics.
One chain, three tariff lines
Sand → oxychloride → oxide spans two chapters and three headings of the nomenclature:
Three headings, three duty treatments in the one schedule we could fully verify. Korean HSK rates and any FTA phase-out we did not verify and do not state; they must be confirmed item by item, as our Korean YSZ import guide concludes one stage downstream.
Finally, a negative finding worth stating plainly. We searched China's April 2025 and October 2025 export-control tranches — the announcements as translated and the analyses of them — for any entry covering zircon, zirconia or zirconium oxychloride. We found none. The only zirconium-bearing controlled item is the yttrium–zirconium alloy sputtering target, controlled for its yttrium, a regime covered by our yttrium export control analysis and the China export controls timeline. Absence of evidence in English-language commentary is not proof of absence: obtain a written 자가판정 or 사전판정 rather than relying on our search.
What to put in the purchase specification
- The route, named — caustic-fusion/ZOC, thermal dissociation (fused), or precipitation. It decides the impurity fingerprint and the radiological status at once.
- The purity convention — ZrO₂ + HfO₂ or ZrO₂ alone? Get HfO₂ as its own line either way.
- Named impurity limits, not one purity number — SiO₂ and Cl⁻ as route markers, Na₂O because the caustic fusion puts it there, plus Fe₂O₃ and Al₂O₃.
- Lot gamma spectrometry in Bq/g for ²³⁸U, ²³²Th and ⁴⁰K, on sand, flour or fused zirconia.
- The country, or mine, of origin — activity is deposit-specific.
- The HS heading, fixed before shipment.
Frequently Asked Questions
Is zirconia radioactive?
There is no single answer, and that is the point. On Selby's figures raw zircon runs 3–4.5 Bq/g of ²³⁸U, chemical-route zirconia below 0.1 Bq/g, and fused zirconia above 1 Bq/g. The route decides, not the compound.
Do we have to register with the NSSC to import zircon sand?
Read literally, 생활주변방사선 안전관리법 제9조 reaches a person who imports or sells 원료물질, and typical zircon exceeds both the 1 Bq/g threshold in 제9조제2항 and — at roughly 231–324 kg a year — the 1,000 kBq quantity threshold in 시행령 제4조제2항. But we could not confirm how the NSSC applies it in practice, finding no public register of handlers and no zircon-specific guidance. Confirm with the Commission or a licensed radiation institution before committing to a schedule.
Does buying Japanese or European powder remove China from our chain?
Not by itself. Tosoh's own published flow starts at zircon sand and passes through ZrOCl₂·8H₂O, as the caustic-fusion route does by definition; changing supplier changes who performs the hydrolysis. We verified this flow for one Japanese maker, Tosoh; for European makers we verified nothing either way, and where any given maker sources its ZOC is worth asking directly.
Is zirconia or zircon sand on China's export-control list?
We searched China's 2025 export-control announcements for any zircon, zirconia or ZOC entry and found none. The only zirconium-bearing controlled item — the yttrium–zirconium sputtering target — is controlled for its yttrium, not its zirconium. Obtain a written 자가판정 or 사전판정 rather than relying on our search.
References (Public Sources)
- Federal Register, Final 2025 List of Critical Minerals, 7 November 2025 — the by-product table.
- USGS, Mineral Commodity Summaries 2026 and 2025 and Minerals Yearbook 2021, "Zirconium and Hafnium" — production, import sources, prices, tariff lines, and the two zirconium-to-hafnium ratios.
- Iluka Resources Limited, Annual Report 2025 and Quarterly Review to 31 March 2026. Its H1 2026 results (18 August 2026) reached us only via trade press.
- J.H. Selby (Zircon Minerals Committee, Richards Bay), "The industrial uses of zircon and zirconia, and the radiological consequences of these uses", IAEA NORM conference paper, 2007 vintage.
- C.A. Hazin and colleagues (CRCN-NE/CNEN), "Determination of ²³⁸U, ²³²Th and ⁴⁰K in Zircon Sand Products from a Processing Plant in Brazil", IRPA.
- Council Directive 2013/59/Euratom, Annexes VI and VII; IAEA SSR-6 (Rev. 1), 2025 edition, para. 107(f) and Appendix; 대한민국 생활주변방사선 안전관리법 (법률 제20141호) 제9조·제11조 and 시행령 (시행 2025-09-30) 제4조.
- 15 CFR Part 774 Supp. No. 1, ECCN 1C234; WCO Harmonized System Chapter 28; US CBP ruling NY I87658 (17 January 2003).
- Tosoh Corporation, Zirconia brochure (undated; retrieved September 2026); Sigma-Aldrich listing for zirconium(IV) oxide, 99.99% trace-metals basis.
- A listed Chinese zirconium-chemicals producer's audited 2025 annual report (filed April 2026), citing China Customs.
Figures reflect public sources at the time of writing; prices and the Korean administrative position can change.
Nami Tech Solutions works project by project on zirconia and zirconium raw materials rather than from standing inventory. Our part of this chain is the unglamorous one: fixing the route and purity convention in the specification before the first quote is compared, reading lot certificates against that document, and settling the HS heading before the material moves.
