hephaestus-mqmqa

A free calculator for slags, salts and alloys. Load a database, then ask what is present at your composition and temperature, or click a phase diagram and read the phases and how much of each. No code to write and nothing to install; everything runs in your browser and nothing is uploaded. Reads ChemSage .dat, Thermo-Calc .tdb, .utdb and .xtdb files.

1 · Load a thermodynamic database

No database loaded yet.
Two dialects, one engine: ChemSage .dat (the FactSage / ChemApp / Thermochimica family; the only format that carries the MQMQA melt model, so slags and salts live here) and Thermo-Calc .tdb (the Thermo-Calc / OpenCalphad / pycalphad family; alloy solid-solution systems). The .utdb unified dialect, a documented open extension of TDB (open specification), holds both at once: one file can carry an alloy, a slag, and a salt together (see the recycling and steelmaking demos). The open XML .xtdb format (Sundman et al., Calphad 90 (2025) 102849) loads here too and opens in the third-generation card below. Every shipped system's sources, modeling judgments, and known limits are in its PROVENANCE file in the open data directory; code and databases live in the GitHub repository. Every file you upload stays as a reload button, coloured by dialect: .dat purple, .tdb teal, .utdb blue.

Ternary phase diagram · FeO–MgO–SiO2assessed

Liquidus projection of the open iron–magnesium–silicate slag database: primary crystallizing phase across the join, with liquidus isotherms. Computed on the refined equilibrium hull of the 3D view and checked against pycalphad; reproduces the topology of Bowen & Schairer 1935. Move over the diagram to read a composition. This is a precomputed showcase of the built-in FeO–MgO–SiO2 database only — it is not generated for other or uploaded systems. For those, use the live ternary section below, which solves whatever database is loaded.

 
Live melt calculator. Click anywhere in the triangle to solve the equilibrium there in your browser, at the section temperature: which phases are present and how much of each, then the liquid's Gibbs energy and its dominant melt species. Every click is added to the list below, numbered to its marker, so you can compare points.

Steel-slag crystallization · CaO–MgO–SiO2 at FeO 5 mol%assessed

Which silicate crystallizes first from the calibrated CaO–FeO–MgO–SiO2 slag melt, and at what temperature. The liquid, the olivine / orthopyroxene / clinopyroxene solid solutions and the oxide solids sit on one calibrated scale (diopside melts at 1673 K, forsterite at 2080 K, 83 K below the measured value); the multi-phase equilibrium is precomputed and validated against pycalphad. Mafic melts crystallize olivine, calcium-rich melts crystallize clinopyroxene. Move over the diagram to read a composition, its primary phase and the liquidus. Like the section above, this is a precomputed showcase of the built-in slag database only, not generated for other or uploaded systems — the live ternary below covers those.

 

Live ternary phase diagramlive engine

The full phase assemblage, solved live in your browser at the temperature you choose — not a precomputed picture. The engine computes the equilibrium liquid across the whole composition field (the WASM MQMQA solver), pools it with the system's solid solutions and stoichiometric compounds, and takes the lower convex hull to find which phases coexist. When you load a database above, its three-cation system appears here automatically and replaces the built-in demos; otherwise pick a demo. Press Compute, then hover to read a composition and its stable phases. The built-in iron–magnesium slag reproduces the assessed section above (validated against pycalphad), and the lime–alumina–silica demo is anchored to 45 published silica activities (Knudsen effusion and gas equilibrium). A live section is a touch coarser at the field boundaries because it is sampled on the fly.

Press Compute to solve the diagram at the chosen temperature.
Reading this diagram. Every phase boundary here is solved live by the open engine, not looked up — the liquid by the MQMQA solver, the solids by their compound-energy models, combined on a lower convex hull. It reproduces the assessed FeO–MgO–SiO2 section (checked against pycalphad); exact boundary positions depend on the sampling grid, so a live section is slightly coarser than the precomputed one above. This system is assessed from published data. For other systems, or if a result looks off, email info@odinzen.io.

3D phase spacelive engine

The liquidus surface over the composition triangle, with height as temperature: each coloured field is where one solid crystallizes first, and the dark lines between fields are the eutectic valleys. Drag to turn it; scroll to zoom. The green plane is a temperature: drag the slider to move it, release, and the exact isothermal section at that temperature is drawn in it. Click any point for the engine's own answer there (its liquidus and the phases present at the plane's temperature); each click is numbered on the plot and saved in the list below. Hovering only reads the drawn surface. Click a legend entry to hide or show a field. Everything is computed in your browser by the open engine, in the background, so the page stays responsive.

Press Start; what is running, its time left, and the queue show here.

Phase diagram builder · build & tune your ownyour data

Enter each component's own data — melting point and heat of fusion, plus a single liquid interaction energy W — and the engine computes the eutectic live from multicomponent regular-solution thermodynamics, for up to four components, solved in your browser. Change any value (the sliders are quickest) and watch the diagram and the eutectic move. Load a preset to start, or enter a brand-new system.

One symmetric W applies to every pair (a first approximation). W < 0 (components attract): a deeper eutectic — the deep-eutectic-solvent regime; strongly negative → compound formation. W > 0 (components repel): a shallower eutectic that rises toward liquid immiscibility as W grows.
 
If you build with your own data, read this. This is a regular-solution model with a single symmetric interaction: it turns your melting points, heats of fusion, and W into a diagram, but it is an approximation and the result is only as good as the numbers you enter. Treat a diagram you built as a starting hypothesis to check against your own data and judgment — it does not replace a proper thermodynamic assessment. For three or more components the eutectic is a numerical estimate over the composition simplex, and one symmetric W will not reproduce a strongly-ordered real system exactly (per-pair interactions and compounds are the assessed-engine path). An unexpected diagram is usually the input data (or a typo), not the engine. Stuck, or think something is off? Email us at info@odinzen.io and we will help you sort it out.

Gas equilibrium · NASA polynomialsgas engine

The ideal-gas equilibrium at a chosen temperature and pressure, solved in your browser from an open NASA-polynomial database (the combustion school's format: NASA CEA, GRI-Mech). Enter how many moles of each element you start with, pick temperature and pressure, and the engine returns the equilibrium gas made of C, H, O, and N species. Validated against Cantera to a part in a million. Tick couple to condensate and the same solve holds the gas and a steelmaking condensate (graphite, wustite FeO, iron) at one shared oxygen potential, so a reducing gas pulls FeO down to iron and an oxidising one holds it as FeO, checked against Cantera's own multiphase equilibrium.

Presets:

Slag ↔ furnace gas · iron-saturatedcoupled

The furnace atmosphere and the melt settle at one oxygen potential. For the assessed FeO-SiO2 slag held at iron saturation, the engine gives the FeO activity, and the gas engine turns it into the CO₂/CO ratio the melt sits under. Add silica and the FeO activity falls, so a more reducing gas is needed to keep iron from oxidising, exactly as slag-gas equilibration is measured. The pure-FeO limit is the iron-wustite buffer, within a percent of Cantera. Drag the silica content to read it off.

0.20

Third-generation diagram · XTDB3rd-gen

A separate reader for the open XML XTDB format (Sundman et al., Calphad 90 (2025) 102849), the format the CALPHAD community is standardising on. It assembles third-generation Gibbs energies down to absolute zero: an Einstein heat-capacity term in place of the old polynomial, and a two-state description of the liquid, so a single function spans cryogenic to superheated temperatures. This is a distinct engine from the ChemSage .dat / Thermo-Calc .tdb calculator above; it reads XTDB files and computes their binary phase diagram live in your browser by common-tangent construction, no server.

Al–C is the third-generation assessment of He et al. (2021); Pb–Sn is a classic assessment written in the XTDB format.

Request a database for your systemconcierge

Have a stack of published papers on a system the open database does not cover yet, whether your own work or literature you have gathered (phase-diagram studies, calorimetry, activity measurements)? That is what an assessed database is built from, the way the open slag database above was made: measured values extracted with provenance, an excess model fitted by this same engine, and a valid database file out the other end — ChemSage .dat when the system needs the MQMQA melt (slags, salts), or Thermo-Calc .tdb for alloy solid-solution systems. This page reads and calculates both.

Published data, opened. No coding is involved at any point: you send papers, and what comes back is a working button on this page or a quoted proposal. If a system is a bounded, natural addition to the open database, we may assess it and add it here as another open database, free, and send you a starter file in the dialect you work in. Sources are cited to their original authors; the contribution is credited to whoever brought it. That covers low-maintenance growth of the commons, and light ongoing additions are welcome. It is not a blanket free service, though: open additions are taken a few at a time as capacity allows, and a large or long-running assessment, or heavy ongoing upkeep, is paid work. We open the science, but sustaining it needs a work product on our side once the scope grows past a small system.

Proprietary or confidential data is assessed privately and the file stays yours, as a paid service. Either way, the same address: point this card at your paper folder to list the papers, then email info@odinzen.io. Your files are listed locally in your browser and never uploaded, there is no server behind this page.

No papers selected yet.