Radial Magnets, Inc.
Radial Magnets, Inc.we know magnets
Radial Magnets, Inc.
Radial Magnets, Inc.we know magnets
Radial Magnets · Technical Resource
Samarium cobalt is the material you move to when neodymium runs out of temperature, stability or corrosion resistance. It costs several times as much per unit of energy and it is more fragile to handle, so it is rarely the first choice — but in hot, corrosive, radiation-exposed or precision-stable applications it is frequently the only choice that works.
“SmCo” covers two distinct alloys with meaningfully different properties, and specifying one when you needed the other is a common and expensive error. Both are sintered rare-earth magnets and both are brittle, but their temperature ceilings differ by 100 °C and their energy products by roughly a third.
| Property | SmCo5 (1:5) | Sm2Co17 (2:17) |
|---|---|---|
| BHmax | 16–25 MGOe | 26–32 MGOe |
| Br | 0.85–1.05 T | 1.05–1.25 T |
| Hcj | 18–30+ kOe | 18–30+ kOe |
| Max operating temperature | 250 °C | 350 °C |
| Curie temperature | ~720 °C | ~820 °C |
| α (Br temp coefficient) | −0.04 %/°C | −0.03 %/°C |
| Corrosion resistance | Excellent, no coating needed | Excellent, no coating needed |
| Cost relative to NdFeB | 3–5× | 4–7× |
| Grade | System | BHmax (MGOe) | Br (kG) | Hcb (kOe) | Hcj (kOe) | Max temp |
|---|---|---|---|---|---|---|
| SmCo 18 | 1:5 | 17–19 | 8.6–9.6 | 8.0–9.0 | ≥18 | 250 °C |
| SmCo 20 | 1:5 | 19–21 | 9.0–9.8 | 8.5–9.3 | ≥20 | 250 °C |
| SmCo 22 | 1:5 | 21–23 | 9.6–10.5 | 9.0–9.8 | ≥20 | 250 °C |
| SmCo 26 | 2:17 | 25–27 | 10.5–11.5 | 9.5–10.5 | ≥22 | 350 °C |
| SmCo 28 | 2:17 | 27–29 | 10.8–11.8 | 10.0–10.8 | ≥22 | 350 °C |
| SmCo 30 | 2:17 | 29–31 | 11.0–12.2 | 10.2–11.0 | ≥22 | 350 °C |
| SmCo 32 | 2:17 | 31–33 | 11.5–12.5 | 10.5–11.5 | ≥20 | 350 °C |
Within the 2:17 family, suppliers offer high-Hcj variants specifically for service near the 350 °C ceiling, and low-temperature-coefficient variants tuned for instrument stability. These are not interchangeable with the standard grade at the same energy product. If the application is genuinely at the temperature limit or genuinely precision-stable, name the requirement — “Hcj ≥ 25 kOe at 300 °C” — rather than a grade number.
Grade tables describe room-temperature values. For any SmCo application worth using SmCo for, the room-temperature numbers are not the design case. Request demagnetisation curves at the actual operating temperature; every reputable mill has them, and the difference between the curve you assumed and the curve you get is where designs fail. The general method is covered in magnets and temperature.
NdFeB outperforms SmCo on energy density and cost at room temperature, and that covers most applications. The cases below are the ones where it does not.
The highest NdFeB grades reach roughly 230 °C, and they get there by loading in dysprosium and terbium, which is expensive and supply-constrained. Sm2Co17 runs to 350 °C as standard. Above about 200 °C in continuous service the cost gap narrows considerably, because you are comparing SmCo against a heavily doped EH or AH grade rather than against an ordinary N42.
This is the reason most often overlooked. SmCo's reversible temperature coefficient of Br is around −0.03 to −0.04 %/°C, against roughly −0.12 %/°C for NdFeB — a factor of three to four. In a sensor, a meter, a travelling-wave tube or any device whose calibration depends on field, that difference is the whole specification. A 50 °C ambient swing moves an NdFeB circuit by about 6 % and an SmCo circuit by under 2 %.
SmCo does not require plating in most industrial, aerospace and marine environments. That removes an entire failure mode — no coating to chip, no edge coverage to argue about, no dimensional stack from plating, and no coating temperature ceiling sitting below the magnet's own. In assemblies that are hard to service, this is worth more than the datasheet suggests.
SmCo is markedly more radiation-tolerant than NdFeB, which is why it dominates in space hardware, accelerator instrumentation and nuclear service. Where NdFeB is used in those environments it generally requires shielding or a service-life derating that SmCo does not.
Properly stabilised SmCo shows very little ageing flux loss over decades. For reference magnets, instrument standards and hardware expected to hold calibration for fifteen or twenty years without recalibration, that predictability is the specification — not the energy product.
SmCo is more brittle than NdFeB, which is already brittle. It chips at edges, cracks under point loading, and is unforgiving of the kind of handling that sintered neodymium tolerates. Most SmCo problems in production are mechanical, not magnetic.
Sm2Co17 needs a substantially higher saturating field than NdFeB — a fixture sized for neodymium will frequently leave SmCo partially magnetised, producing parts that measure low and behave inconsistently. If you magnetise in-house or at final assembly, confirm the fixture's peak field against the grade before committing to the design. Under-magnetisation is a common root cause of “the SmCo is weaker than the datasheet” complaints.
Practical minimum wall and thickness dimensions are larger for SmCo than for NdFeB, and yields fall steeply as sections get thin. Where a design calls for a very thin SmCo section, discuss it at RFQ stage — the part may be manufacturable but at a scrap rate that changes the economics.
SmCo's commercial profile is genuinely different from NdFeB's and it needs planning for rather than discovering.
Comparing SmCo and NdFeB on price per kilogram is misleading in both directions. The design comparison is total delivered cost for a magnet that meets the field requirement at the maximum operating temperature — and at high temperature the NdFeB option is a doped EH or AH grade that has itself become expensive, in a larger size than the room-temperature calculation suggested. Run both options to a working geometry at temperature before deciding. It is not unusual for SmCo to lose on paper and win on the built assembly.
Where the driver for considering SmCo is rare-earth supply exposure rather than temperature, the trade-offs run differently — see rare-earth-free and reduced-Dy options and supply risk and export controls.
We supply Sm2Co17 and SmCo5 in disc, ring, block, arc and custom geometries, and we hold buffer stock on the common grades. Send us the operating temperature, the field requirement and the envelope and we will work back to a grade and geometry.