Magnetocaloric materials.
Cooling with a magnetic field.
A magnetocaloric campaign finds refrigerant materials that warm and cool in a magnetic field, for refrigerators and heat pumps that do not compress a gas.


- Magnetic refrigeration
- Heat pumps
- Gas liquefaction
Why this domain matters to your work.
Magnetocaloric materials warm when a magnetic field is applied and cool when it is removed, which lets a refrigerator or heat pump move heat without compressing a gas. The effect peaks near a magnetic transition, so the device’s working temperature has to sit close to the material’s transition. Gadolinium is the reference refrigerant near room temperature, and first-order families such as Gd₅(Si,Ge)₄, La(Fe,Si)₁₃ and (Mn,Fe)₂(P,Si) offer larger effects for magnetic refrigeration, heat pumps and gas liquefaction.
The largest effects come from first-order transitions, and those usually bring hysteresis, which wastes work in every cycle and limits how fast a device can run. Second-order refrigerants cycle reversibly but give a smaller entropy change. The transition also has to land at the right temperature, and the cooling power and adiabatic temperature change must be useful in the modest fields that permanent magnets supply. Gadolinium and the other rare earths carry cost and supply risk, which is why iron-rich and rare-earth-free families draw attention.
A campaign lets your team name the working temperature, the size of effect you need, the hysteresis you can accept, and your limits on rare earths and cost. You receive ranked gadolinium, Gd₅(Si,Ge)₄, La(Fe,Si)₁₃, (Mn,Fe)₂(P,Si) and Heusler candidates, each with estimated transition temperature and order, entropy change at 2 T, adiabatic temperature change, hysteresis, cooling power and cost. These are screening estimates, not calculations for the new composition. Fields above 2 T, hydrogenated La(Fe,Si)₁₃, cycling fatigue, corrosion and regenerator performance are decided in the lab and the prototype.
The material families this domain searches.
Each family brings its own structures and chemistry to the search. A campaign can search one of them or compare several under the same targets.
Gadolinium solid solutions
Gadolinium, the reference refrigerant near room temperature, alloyed to tune its transition, chosen as a reversible and well-understood starting point.
Gd₅(Si,Ge)₄
Gadolinium silicide-germanides with the giant effect of a first-order transition, chosen when the largest effect matters more than cost.
La(Fe,Si)₁₃
Lanthanum–iron–silicon compounds that are mostly iron, chosen for a large effect with far less rare earth than gadolinium refrigerants.
(Mn,Fe)₂(P,Si)
Manganese–iron phosphide-silicides with a first-order transition and no rare earth, chosen for refrigerants made from plentiful elements.
Ni–Mn–X Heuslers
Heusler alloys that cool when a field is applied, an inverse effect of a combined magnetic and structural transition, explored for rare-earth-free devices.
The properties a campaign can aim for.
Choose the properties that matter to your work. For each one, say whether a candidate must meet it or should push it as far as it goes, with cost and supply always in view.
- Entropy change in a field
- Transition near room temperature
- Low hysteresis
- Relative cooling power
- Adiabatic temperature change
- Cost and supply risk
Starting points for a campaign.
Each starting point sets up a campaign for a common goal in this domain. Begin from one and adjust its targets, or set your own from the start.
Room-temperature refrigerants
Suits refrigerator and heat-pump designers, holding the transition between 250 and 350 K so the material works near room temperature.
Giant magnetocaloric effect
Suits teams that need the biggest effect per cycle, searching the first-order families for the largest entropy change.
Low-hysteresis second-order refrigerants
Suits fast-cycling devices, searching for reversible refrigerants that waste little work in each cycle.
Rare-earth-free refrigerants
Suits teams avoiding rare-earth supply risk, searching for refrigerants with no rare earth on a tight budget.
What you specify, what you receive and where the figures stop.
MS Lattice hands your lab a shortlist to make and test. Its figures are screening estimates and calculations, not measurements of the new material, and each domain says where they stop.
You specify
The working temperature, the size of effect you need, the hysteresis you can accept, and your limits on rare earths and cost.
You receive
Ranked gadolinium, Gd₅(Si,Ge)₄, La(Fe,Si)₁₃, (Mn,Fe)₂(P,Si) and Heusler candidates, with estimated transition temperature and order, entropy change at 2 T, adiabatic temperature change, hysteresis, cooling power and cost.
Before you rely on it
The magnetocaloric figures are screening estimates, not calculations for the new composition. Fields above 2 T, hydrogenated La(Fe,Si)₁₃, cycling fatigue, corrosion and regenerator performance are not covered.
The same four modes work in every domain.
Search from your targets alone, improve a material your team already uses, or watch the engine reach known materials, with the targets as guides or sealed away, to see how well it searches.
Open Discovery
Choose it when you know what the material must do but have no starting structure.
Prototype-Guided Optimization
Choose it when you already have a good material and want to make it better.
Guided Rediscovery
Choose it when you want to watch the engine reach a material you already know.
Blind Rediscovery Benchmark
Choose it when you want to test whether the engine finds an answer it was never shown.
Every candidate arrives with its record.
Each candidate comes ranked against your targets, with its estimated properties and their uncertainty, the checks it passed and a structure file to export, ready for your lab.
Questions teams ask about this domain.
Have a materials target? Tell us what it must do.
Share the material domain, the properties you need and the limits the material must respect. We will show you how MS Lattice would approach it through Norck, from a ranked shortlist for your own lab all the way to finished parts.

