BRENNEN

Industries / Mining and materials / Mining and quarrying / Rare earth processing

Radioactive waste

Rare earth minerals carry thorium and uranium, and this work is the storing, treating and cleaning up of the radioactive residues that processing them leaves behind.

5

Milestones achieved, from 1991 to 2019.

1

To watch for: announced, not yet confirmed outside.

3

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1991

    Taking down a monazite plant

    Monazite, a sand mineral that the International Atomic Energy Agency (IAEA) says typically holds 5 to 6% thorium oxide, was treated for rare earths at Udyogamandal, India, without a break from 1952 to 1988. The plant was decommissioned over nine months in 1990 and 1991. Radium had soaked into the rubber linings of the reaction tanks, which measured 55,500 becquerels per gram (Bq/g, a measure of radioactivity in a material), more than four times the radium-bearing cake the plant made on purpose. At an Indian monazite plant, thorium hydroxide was stored in concrete silos 30 to 90 cm thick, over which gamma readings matched local natural background, the IAEA reported in 2011 from the company's own measurements.

    LessonRadioactivity gathers where nobody plans to store it, in tank linings and pipework, so a plant's equipment becomes waste too.

    www-pub.iaea.org

    Credit

    Indian Rare Earths Limited (IREL); IAEA (2011 report)

  2. 1992

    Closed by its own waste

    Asian Rare Earth Sdn Bhd (ARE), owned by Mitsubishi Chemical, began extracting yttrium from monazite near Bukit Merah, in Perak, Malaysia, in July 1982. It shut in 1992 under local and international pressure over its radioactive waste. Residents say they have suffered high numbers of birth defects and leukemia. The waste went to a long-term store at Kledang, completed in September 1988. Retrieval of the drums and building of sealed cells to hold the contaminated material began only in 2010, and in 2019 the Perak government's corporation said the buried waste met regulatory requirements and kept a 200 m exclusion zone around it.

    LessonClosing a plant does not end its waste problem; the cleanup began 18 years after this one closed.

    themalaysianinsight.comeco-business.com

    Credit

    Asian Rare Earth Sdn Bhd (Mitsubishi Chemical); Perak state government (cleanup)

  3. 1998

    Leaks stop a separation plant

    Mountain Pass in California, mined for rare earths since 1952, piped its waste salt water to evaporation ponds on the Ivanpah dry lake. In 1998 its owner, Unocal, suspended all chemical processing, mainly because of leaks in that pipeline, which released water carrying salt and small amounts of dissolved rare earths and radioactive elements. Mining stopped in 2002 for low prices, a lack of room for tailings (the crushed waste left after the ore is treated) and delayed permits for a new tailings store. The next owner, Molycorp, said in its 2010 share offering that it would instead turn waste salt water back into hydrochloric acid and sodium hydroxide for reuse.

    LessonA mine's waste lines and tailings space can halt production as surely as prices can.

    sec.gov

    Credit

    Unocal (operator); Molycorp (2010 filing)

  4. 2017

    Thorium kept, not discarded

    France's national plan for radioactive materials and waste required Solvay to report its thorium holdings to the energy minister by the end of 2017. Its report counts 6,247 tonnes of thorium, about 21,200 m³ once packed, at its Rhodia site in La Rochelle, left from treating monazite: crude thorium hydroxide made from 1970 to 1987 and thorium nitrate made from 1978 to 1994. France classes these as materials with a possible future use, not as waste, and the report sets out how they could be treated and placed in a low-level disposal site if that changes.

    LessonCalling a residue a future resource postpones, but does not remove, the need for a place to put it.

    asn.fr

    Credit

    Solvay (Rhodia); French energy ministry

  5. 2019

    The residue stays in Malaysia

    Lynas separates rare earths at Gebeng, near Kuantan, Malaysia, and its processing leaves a thorium-bearing by-product called water leach purification (WLP) residue. In August 2019 Malaysia dropped its demand that Lynas ship 450,000 tonnes of it out of the country, after Australian authorities refused to accept it, and required a permanent disposal facility instead. In November 2024 Lynas opened a plant at Kalgoorlie, Western Australia, that does the cracking and leaching (breaking down and dissolving the concentrate) and ships an intermediate to Malaysia for separation. On 18 November 2026 Malaysia's deputy science minister told Parliament the disposal facility was 79% built as of 31 October and held 1.46 million tonnes of WLP residue.

    LessonA plan to ship the residue abroad failed when Australia refused to take it.

    malaymail.comsmallcaps.com.aubernama.com

    Credit

    Government of Malaysia (Ministry of Science, Technology and Innovation; Atomic Energy Department); Lynas Rare Earths

Watch for

  1. Thorium out of the residue

    Malaysia renewed Lynas's license on 3 March 2026 for ten years, on condition that it stops producing WLP residue by 2031 and treats what it makes until then to below 1 Bq/g, the level at which it counts as ordinary scheduled waste rather than radioactive waste. The science minister said a research project completed in 2025 had shown "very promising results" in extracting thorium from the residue at laboratory scale, and that Lynas had committed to the timeline.

    Achieved whenMalaysia's Atomic Energy Department reports residue from the full-scale plant measured below 1 Bq/g, or reports that WLP production has stopped.

    thestar.com.mymalaymail.com

    Claimed by

    Chang Lih Kang, Malaysia's Minister of Science, Technology and Innovation, for Lynas, March 2026

Open

  1. 01

    Archive Malaysia's radiation readings

    A public archive of the gamma dose-rate readings that Malaysia's Atomic Energy Department posts live on its Environmental Radiation Monitoring System (ERMS) website, which the government points to for the Lynas plant at Gebeng. Save every station's 10-minute reading as it appears, and publish monthly summaries against the 0.5 microsieverts per hour line that the site itself marks as "Bahaya" (danger).

    The brief
    Credit · unclaimed
  2. 02

    A ledger of stored thorium

    A public table of every known store of thorium-bearing residue left by rare earth processing: site, owner, years produced, tonnes of residue and of thorium, chemical form, how it is stored, and whether the regulator classes it as waste or as material kept for future use, each figure with a dated public source.

    The brief
    Credit · unclaimed
  3. 03

    Baotou's tailings pond from space

    A yearly map, from 1986 to 2026, of the tailings pond near Baotou in Inner Mongolia, China, that holds the waste from processing Bayan Obo ore, which supplied nearly half the world's rare earths by the IAEA's 2011 account: its outline, open water, wet and dry tailings surface, and the land use within 2 km, from free Landsat and Sentinel-2 satellite images, with an area table for each year.

    The brief
    Credit · unclaimed

Open 01 · the brief

Archive Malaysia's radiation readings

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public archive of the gamma dose-rate readings that Malaysia's Atomic Energy Department posts live on its Environmental Radiation Monitoring System (ERMS) website, which the government points to for the Lynas plant at Gebeng. Save every station's 10-minute reading as it appears, and publish monthly summaries against the 0.5 microsieverts per hour line that the site itself marks as "Bahaya" (danger).

Why it is still open

The ERMS page shows only current 10-minute and 60-minute readings and its thresholds, with no history or download. The deputy science minister told Parliament on 18 November 2026 that the public can follow the data "in real-time". Searches of the web and GitHub found no independent archive or analysis of the readings.

erms.atom.gov.mybernama.com

Built with what exists

Python with Playwright to read the page, which loads its data by script; a scheduled job on GitHub Actions or a small cloud server polling once every 10 minutes; SQLite or Parquet files published on GitHub or Zenodo; a plotting library such as Matplotlib.

  1. First step

    Open the ERMS page with a browser's developer tools, find the request that delivers station readings, and save one full day of them.

  2. Done looks like

    Twelve months of readings from every listed station, with each gap logged, published with a monthly table of median and maximum readings per station and a count of readings at or above 0.5 microsieverts per hour.

Open 02 · the brief

A ledger of stored thorium

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A public table of every known store of thorium-bearing residue left by rare earth processing: site, owner, years produced, tonnes of residue and of thorium, chemical form, how it is stored, and whether the regulator classes it as waste or as material kept for future use, each figure with a dated public source.

Why it is still open

The figures sit in scattered documents. Solvay's 2017 report gives 6,247 t of thorium at La Rochelle, and Malaysia's ministry gives 1.46 million t of residue moved into Lynas's disposal facility. The IAEA's 2019 survey of thorium resources mentions some stores, such as 4,000 t of thorium cake in 20,000 drums in Brazil, but as notes on possible future supply, not as a register of stores. The IAEA's 2011 safety report describes residue handling plant by plant with no world tally. A 2020 review by Aziman and Ismail treats rare earth residue as a thorium resource; whether it lists stores site by site was not checked. Searches found no site-by-site register.

asn.frbernama.comwww-pub.iaea.orgwww-pub.iaea.orgapi.crossref.org

Built with what exists

Public reports from the IAEA, national radioactive waste inventories, company filings and parliamentary answers; Python with pdfplumber for tables; an AI model such as Claude to pull candidate figures out of scanned reports, with every figure checked by hand against its page (the Note "Seven more flagged" covers a Claude-based document reviewer).

  1. First step

    Enter the five sites in this section (Udyogamandal, Bukit Merah, Mountain Pass, La Rochelle and Gebeng) and the Brazilian drums, then search national waste inventories for other holders.

  2. Done looks like

    At least 15 sites in at least six countries, every number linked to the page it came from, the La Rochelle and Lynas figures reproduced exactly, published as a CSV file under an open license.

Open 03 · the brief

Baotou's tailings pond from space

Credit · unclaimed

The first to publish is credited on this line, and the deliverable moves from Open to Achieved.

What to build

A yearly map, from 1986 to 2026, of the tailings pond near Baotou in Inner Mongolia, China, that holds the waste from processing Bayan Obo ore, which supplied nearly half the world's rare earths by the IAEA's 2011 account: its outline, open water, wet and dry tailings surface, and the land use within 2 km, from free Landsat and Sentinel-2 satellite images, with an area table for each year.

Why it is still open

The IAEA describes the tailings being pumped to a pond and kept wet to hold down dust, from ore with about 0.04% thorium oxide, and BBC Future described its shore lined with waste pipes in 2015. Published satellite studies found in searches cover the clay rare earth mines of southern China, such as a 2016 study of three mining areas in Jiangxi from 2000 to 2010, not the Baotou pond. A personal project on GitHub samples Sentinel-2 moisture, bare-soil and vegetation indices over the pond and compares its winter snow cover in 2024 and 2025, but publishes no outline, no yearly areas and no accuracy check. Chinese-language databases were not searched.

www-pub.iaea.orgbbc.comijetch.organdjames.github.io

Built with what exists

Google Earth Engine or Microsoft Planetary Computer; the free Landsat and Sentinel-2 image archives; a water index and a supervised classifier such as random forest; hand-checked sample points on high-resolution imagery.

  1. First step

    Draw the pond's outline on a 2025 Sentinel-2 image and measure its area, then repeat on a 1990 Landsat image.

  2. Done looks like

    A map and an area figure for every year from 1986 to 2026, with a published accuracy check against at least 200 hand-labeled points, and the code and outlines released.