BRENNEN

Industries / Mining and materials / Glass and ceramics / Refractories

Bricks and shapes

Bricks and shapes are linings made before they reach the furnace: grains of a heat-resistant mineral are mixed with a binder, pressed in molds and then fired in a kiln or simply cured, as standard bricks or as special parts such as the nozzles and sliding plates that control the flow of molten steel.

4

Milestones achieved, from 1964 to 1978.

1

To watch for: announced, not yet confirmed outside.

2

Open deliverables. Nobody has shipped them.

The road

AchievedWatch forOpen

Achieved

  1. 1964

    A sliding gate pours steel

    A sliding gate controls the stream of molten steel leaving a ladle by sliding one refractory plate across another under the outlet. David D. Lewis of Johnstown, Pennsylvania, patented a gate of this kind, holding perforated refractory bricks pressed together by springs, in February 1885 (US 311,902). By the account of Krosaki Harima, a Japanese refractory maker, the idea was first put to practical use in Germany in 1964 and spread worldwide from about 1969. Yawata Steel Works used one on a 30-tonne ladle in 1968; Krosaki delivered its first units in 1970, applied them to tundishes (the troughs that feed a continuous caster) in 1972, and had sold 4,000 by 1993. The plates changed from carbon-free zircon and high-alumina grades to alumina-carbon around 1975.

    krosaki.co.jppatents.google.com

    Credit

    David D. Lewis; Yawata Steel Works; Krosaki Harima

  2. 1976

    Resin binds magnesia and graphite

    Magnesia-carbon (MgO-C) bricks mix magnesia grains with flaky graphite and are cured, not fired. Nippon Steel's engineers call them "epoch-making": the graphite improved resistance to slag and, because it expands little, absorbed the expansion of the magnesia and cut spalling (cracking and flaking from heating and cooling), two needs that had long pulled against each other. Krosaki Harima says its Shimizu plant put unfired MgO-C bricks, developed in 1971, into electric arc furnaces, where they outlasted magnesia-chrome bricks, but had to stop mass production because the odor of the phenol resin binder fouled the mixing and molding rooms. After trial runs with more than 250 binders made with a resin supplier, a liquid novolac resin was developed in 1976. The resin-bonded bricks, with water-cooled furnace walls, changed how electric arc furnaces were lined.

    LessonA binder has to suit the people who mix and press it as well as the furnace.

    krosaki.co.jpnipponsteel.com

    Credit

    Krosaki Harima's Shimizu plant and an unnamed binder maker

  3. 1977

    Graphite bricks in converter bottoms

    In 1977 Kawasaki Steel's Chiba Works (now part of JFE Steel) brought in Japan's first Q-BOP, a converter that blows pure oxygen through tuyeres (nozzles) in its bottom. Its fired dolomite bottom bricks lasted only 200 to 300 heats, a heat being one batch of steel. Yomoda, manager of Krosaki Harima's Shimizu plant, proposed MgO-C bricks with 20% graphite, which others in the field thought dangerous so close to the oxygen tuyeres. Trials began at the tuyeres in April 1977 and across the bottom in August. The bottom lasted three times as long as with dolomite, and after changes to installation and smaller tuyere bricks it passed 2,000 heats within about three years, by Krosaki's account. MgO-C bricks then spread to every part of the converter; Nippon Steel's 2013 review says they became widely used in converters in the 1980s.

    LessonA full-scale trial overturned a shared expert belief about what the material could survive.

    krosaki.co.jpnipponsteel.com

    Credit

    Yomoda of Krosaki Harima's Shimizu plant; Kawasaki Steel, Chiba Works

  4. 1978

    Nozzles pressed from all sides

    Continuous casting turns molten steel straight into slabs or billets instead of ingots, and it needs shapes that ingot casting did not: sliding-gate plates, ladle shrouds and submerged entry nozzles, tubes that carry the steel below the surface in the mold to keep air off it. Krosaki Harima says alumina-graphite nozzles reached Japan from an overseas maker in 1970; it installed a small cold isostatic press, which squeezes powder in a flexible mold from all sides with fluid pressure, in 1971, and began full-scale production in 1978. Nippon Steel notes that inclusions (particles of oxide carried in the steel) have stuck to and clogged these nozzles since they came into use in the 1970s. The shift hit older products: clay "surface bricks" used in large amounts for ingot casting declined from 1970, and Krosaki says its refractory output fell from a peak of 470,000 tonnes a year to just under 160,000 tonnes in 1993.

    LessonWhen steelmakers changed how they cast, a large market for plain bricks shrank and a smaller one for precision shapes grew.

    krosaki.co.jpnipponsteel.com

    Credit

    Krosaki Harima

Watch for

  1. Low-carbon bricks in degassers

    Krosaki Harima says its MAGreenX bricks, unfired magnesia-carbon bricks with less carbon than usual, are replacing magnesia-chrome bricks in the lower vessels and snorkels of RH degassers, vessels that draw gases out of molten steel under vacuum. It says one customer reported 1.5 times the performance after switching, that its own simulations show the degasser's outer surface running tens of degrees cooler than with ordinary MgO-C bricks, and that making a tonne of unfired brick emits about 80% less CO2 than a fired brick. Its sales manager adds that not every customer will see the same effect.

    Achieved whena steelmaker or outside party publishes lining lives and shell temperatures for RH vessels lined with these bricks, compared with magnesia-chrome linings at the same plant.

    krosaki.co.jpkrosaki.co.jp

    Claimed by

    Krosaki Harima, 2026 (the pages are undated; their images were posted in January 2026)

Open

  1. 01

    Open data for carbon bricks

    An open table of laboratory measurements on magnesia-carbon bricks taken from published papers. For each brick: its recipe (magnesia purity and grain size, graphite and resin content, antioxidants such as aluminum or silicon powder), how it was cured or fired, and what was measured (open porosity, density, cold and hot strength, oxidation loss, slag attack, strength left after thermal shock), with the paper's DOI and table number on every row. Add a baseline model scored on papers it has not seen.

    The brief
    Credit · unclaimed
  2. 02

    Carbon footprints, brick by brick

    A public table of the published carbon footprints of refractory products, drawn from makers' technical data sheets and from environmental product declarations (EPDs, audited statements of a product's life-cycle impacts). For each: product, maker, type (fired brick, unfired brick, castable, fused-cast block), main raw materials, declared unit, which method and life stages are counted, kilograms of CO2 equivalent per tonne, verifier and link. Add a short check of each published percentage claim against the table.

    The brief
    Credit · unclaimed

Open 01 · the brief

Open data for carbon bricks

Credit · unclaimed

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

What to build

An open table of laboratory measurements on magnesia-carbon bricks taken from published papers. For each brick: its recipe (magnesia purity and grain size, graphite and resin content, antioxidants such as aluminum or silicon powder), how it was cured or fired, and what was measured (open porosity, density, cold and hot strength, oxidation loss, slag attack, strength left after thermal shock), with the paper's DOI and table number on every row. Add a baseline model scored on papers it has not seen.

Why it is still open

A 2023 review of machine learning on MgO-C bricks by Sado, Jastrzębska, Zelik and Szczerba found that reliable models need high-quality data in amounts that do not yet exist, calling that "the greatest current challenge" and asking industry to share data. A GitHub project posted in November 2025 trains a random-forest model on 600 MgO-C records but does not say where the records came from. Searches of GitHub, Hugging Face and Zenodo in November 2026 found no pooled table with a source on each row.

ebi.ac.ukgithub.com

Built with what exists

Python with pandas and scikit-learn; WebPlotDigitizer for values printed only in charts; an AI model to find papers and pull their tables, every value checked by hand against the paper (see the Note "Seven more flagged"); Zenodo for a versioned release with a DOI.

  1. First step

    Take the studies cited in the Sado review, open those that are free to read, and tabulate the recipes and results of the first ten.

  2. Done looks like

    At least 400 bricks from at least 40 papers in one openly licensed file, each row traceable to a page; for open porosity and oxidation loss, a published error when each paper in turn is left out of training.

Open 02 · the brief

Carbon footprints, brick by brick

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 the published carbon footprints of refractory products, drawn from makers' technical data sheets and from environmental product declarations (EPDs, audited statements of a product's life-cycle impacts). For each: product, maker, type (fired brick, unfired brick, castable, fused-cast block), main raw materials, declared unit, which method and life stages are counted, kilograms of CO2 equivalent per tonne, verifier and link. Add a short check of each published percentage claim against the table.

Why it is still open

The rules now exist, but the numbers sit in separate documents. RHI Magnesita said in April 2022 that it prints a cradle-to-gate footprint, calculated under ISO 14067, on the technical data sheet of every product (2.211 t CO2e per tonne for one fired magnesia-spinel brick, ANKRAL R1). Eleven companies wrote a common footprint method for refractories within the World Refractories Association from January to November 2024, and Calderys said in April 2025 that it would apply it across its range by the end of that year so that customers could compare products. Saint-Gobain's 2023 EPD for fused-cast furnace blocks (S-P-08602) gives 1.45 kg CO2e per kg, as listed in the One Click LCA database, and Krosaki Harima says an unfired brick emits about 80% less CO2 to make than a fired one, without publishing the figures behind it. A 2019 comparison of two ladle bricks by Francesco Boenzi and colleagues warned that how many bricks a tonne of steel consumes must also be considered. Searches in November 2026 found no table that sets products from different makers side by side.

rhimagnesita.comcalderys.commaterials.oneclicklca.comkrosaki.co.jpiris.poliba.it

Built with what exists

Makers' technical data sheets; the International EPD System's library and other EPD registries; makers' sustainability reports; an AI document reader to pull declared units and figures out of data sheets and EPD files, every number checked by hand (see the Note "Seven more flagged"); a spreadsheet or Python with pandas.

  1. First step

    Search the International EPD System's library for "refractory" and log every declaration it holds.

  2. Done looks like

    At least 50 products from at least 8 makers, each with declared unit, method, life stages counted and a link, plus the published range for fired bricks, unfired bricks and castables; at least three published percentage claims each marked as supported, contradicted or not checkable.