SmCo Magnets

Specification

Alloy families

Sm1Co5 (1:5, YX/XG-16/18/20 class) and Sm2Co17 (2:17, YXG-24/26/28/30/32 class)

Grade range

SmCo16-SmCo20 (1:5); SmCo24, SmCo26, SmCo28, SmCo30, SmCo32 (2:17)

BHmax

16-20 MGOe (127-159 kJ/m3) for 1:5; 24-32 MGOe (191-255 kJ/m3) for 2:17. SmCo32 sits at the practical ceiling of production sintering

Remanence (Br)

0.82-0.95 T (1:5); 1.00-1.15 T (2:17)

Intrinsic coercivity (Hcj)

1195-2000 kA/m (15-25 kOe) typical; 1:5 can reach >2400 kA/m (30 kOe) in low-Br grades

Max operating temperature

250degC standard for both 1:5 and 2:17; 300degC and 350degC 2:17 variants available at reduced Br. Curie point ~800-850degC (2:17), ~720-750degC (1:5). Max service temp is load-line dependent - a Pc of 2 or better is assumed

Temperature coefficients

Br: -0.030 to -0.035%/degC (2:17), -0.04 to -0.045%/degC (1:5). Hcj: -0.2 to -0.3%/degC, with 1:5 toward the steep end. NdFeB is roughly -0.11 to -0.12%/degC Br for comparison

Tolerance and coatings

As-sintered +/-0.1 mm typical; ground +/-0.05 mm, +/-0.02 mm achievable on critical faces. Normally supplied bare (passive oxide is stable to ~300degC in air). Ni, Ni-Cu-Ni, epoxy or parylene only where particulate containment or bonding demands it

Typical applications

• Aerospace and defence actuators, sensors and travelling-wave tube focusing structures, where flux drift over -55 to +200degC must stay within a fraction of a percent
• High-speed and high-temperature PM motors and generators - turbochargers, aircraft engine accessories, downhole tooling
• Precision instrumentation: Hall-effect and magnetoresistive sensors, flowmeters, load cells, where the temperature coefficient dominates the error budget
• Magnetic couplings and pumps handling hot or aggressive fluids, where an NdFeB coating failure would liberate corrosion product into the process
• Vacuum and UHV assemblies - ion pumps, mass spectrometers, particle beam optics - where low outgassing and no plating are mandatory
• Oil and gas MWD/LWD tooling operating continuously at 200degC+ with thermal cycling

Engineering notes

• Cost is the main barrier. Cobalt dominates the bill of materials - roughly 66 wt% of a 1:5 alloy and about 50 wt% of a 2:17 (the balance including 15-20% Fe, 5-6% Cu and 2-3% Zr) - and samarium carries its own supply dynamics. Expect SmCo to run several times the price per kilogram of an equivalent NdFeB part, and the gap widens on a per-MGOe basis. If your worst-case hot-spot is 150degC or below, a high-Hcj NdFeB grade will almost always be the cheaper correct answer.
• It is brittle and weak in tension - roughly 35 MPa tensile, 800 MPa compressive, fracture toughness around 1 MPa*m^0.5, with 2:17 more prone to chipping than 1:5. Thin sections, sharp internal corners, blind threads and unsupported spans are asking for cracks. Keep minimum wall above ~1.5 mm, chamfer all edges, and never treat the magnet as a structural member. Handling and press-fit assembly cause more field failures than the magnetics ever do.
• 2:17 alloys need a controlled solution-treat plus isothermal ageing cycle to develop the cellular precipitate structure that carries the coercivity. Properties are therefore sensitive to section thickness and process discipline, and grade-to-grade Hcj scatter is wider than in NdFeB. Specify Hcj at your actual operating temperature, not just at 20degC, and ask for a demagnetisation curve at temperature.
• Corrosion resistance is good, not universal. The oxide passivates in air and most neutral environments, but SmCo is attacked by strong acids, oxidises measurably above ~350degC in prolonged exposure, and is not immune to hydrogen - avoid pressurised hydrogen service. Uncoated parts also shed fine dark particles under abrasion, which matters in clean optical or bearing assemblies.
• Shaping is abrasive and electrical, not chip-forming: diamond wheel grinding with coolant, plus wire or sinker EDM for slots, holes and profiles (SmCo is conductive). No turning, drilling or tapping - so tolerance and shape complexity drive price more steeply than material volume does. Saturation also demands a higher applied field than NdFeB: plan for roughly 2400-3200 kA/m (30-40 kOe) for 2:17 and 3200-4000 kA/m (40-50 kOe) for 1:5, which is the hardest commercial magnet family to saturate. Magnetising fixtures sized for NdFeB leave SmCo partially magnetised, which shows up as a quiet 5-10% flux shortfall rather than an obvious failure.