engineering guide — quality & reliability

Why Magnets Fail: A Root-Cause Field Guide

Permanent magnets don't "wear out" — properly specified, an NdFeB magnet loses well under 1% of its flux over a decade. When a magnet fails in the field, it failed for a specific, diagnosable reason. This guide covers the four failure families, how to tell them apart, and how to spec them out of your design.

failure mode 1 — thermal demagnetization

Every magnet grade has a maximum operating temperature, but the real story is on the demagnetization curve. As temperature rises, the intrinsic coercivity H_cj of NdFeB drops sharply (β ≈ −0.5 to −0.7 %/°C). The "knee" of the B-H curve moves up toward the operating point — and once the operating point crosses the knee, flux loss becomes irreversible. Cooling the magnet back down does not restore it; only re-magnetization does.

heat raises the knee of the intrinsic curve. if the load line intersects below the knee, flux loss is permanent.

symptom — uniform flux loss across the part, no visible damage, often after a thermal excursion (soldering, paint bake, brake heat, stalled motor)

confirm — measure open-circuit flux vs. a golden sample; loss recovers only after re-magnetization

fix — higher-H_cj grade (M/H/SH/UH/EH suffix), higher permeance coefficient (thicker magnet, smaller air gap), or manage the thermal environment

failure mode 2 — reverse-field demagnetization

External opposing fields do the same damage heat does. Common culprits: stator currents during a locked-rotor or short-circuit event in a motor, adjacent magnets in repelling arrangements (Halbach assemblies, repulsion bearings), magnetizing/handling equipment, and welding near assembled magnets.

The failure is often local — the corner or edge of the magnet nearest the opposing field demagnetizes first, producing a distorted field pattern rather than uniform loss. A flux map (or even a simple gaussmeter scan across the pole face) reveals the signature dead zone.

motor designers: always run the demag check at the worst case — maximum stator current at maximum rotor temperature, simultaneously. It's the combination that kills. A grade that survives either condition alone can fail when both stack.

failure mode 3 — corrosion & coating failure

Sintered NdFeB is a powder-metallurgy material with a neodymium-rich grain boundary phase that corrodes aggressively — and it's porous enough for moisture to work inward. Corrosion failures show up as coating blisters, white/gray powder (neodymium oxide/hydroxide), edge chipping that starts at coating breaks, and in humid+warm environments, hydrogen decrepitation: the material absorbs hydrogen, the lattice swells, and the magnet literally crumbles from the surface.

Where coatings actually fail

edges & corners — plating is thinnest here; chips during assembly expose bare material

bond lines — aggressive adhesive primers or acidic cure byproducts can attack Ni plating

coating porosity — single-layer Ni has pinholes; Ni-Cu-Ni exists precisely to break up through-porosity

wrong coating for the job — Ni-Cu-Ni is not a salt-spray coating. marine and autoclave duty needs epoxy over plating, everbright/parylene, or a switch to SmCo

Diagnosis is straightforward: sectioning and SEM/EDS shows corrosion product chemistry, and salt-spray testing per ASTM B117 benchmarks replacement coating systems. If the environment can't be controlled, remember SmCo and ferrite are essentially corrosion-immune and may be the cheaper system answer.

failure mode 4 — mechanical fracture

Sintered magnets are ceramics in disguise: strong in compression, weak in tension and bending, and intolerant of impact. NdFeB flexural strength is roughly 250 MPa with essentially zero ductility. Typical fracture scenarios:

slam-together impact — two strong magnets or magnet-to-steel collision during handling; chips at edges, sometimes clean face fracture

clamping stress — screws or press fits loading the magnet in tension/bending; sintered magnets should never be primary structural members

thermal shock — rapid temperature swings across a bonded interface with CTE mismatch crack the magnet or the bond

fatigue at cracks — a chip from assembly becomes a crack initiation site under vibration; parts that passed EOL test fail months later

Fracture surfaces tell the story: a fast overload fracture is clean and grainy; a corrosion-assisted fracture shows discolored, powdery regions at the initiation site.

quick diagnostic table

observation	most likely cause	verify with
uniform flux loss, no visible damage, after heat event	thermal demagnetization	flux compare vs. golden sample; re-magnetize and re-test
localized dead zone on one edge/pole	reverse-field demag	gaussmeter scan / flux map of pole face
white-gray powder, blistered or flaking coating	corrosion / hydrogen attack	SEM/EDS on corrosion product; humidity history
chips at edges, cracked or shattered part	mechanical impact or clamping stress	fractography; assembly process review
gradual flux drift over months, warm environment	marginal grade — operating near the knee	elevated-temp flux aging test on samples
works at room temp, fails hot in application	combined temp + reverse field crossing knee	worst-case demag simulation (FEA) at temp

designing failures out

spec H_cj at temperature — pick the coercivity suffix from worst-case temperature + worst-case reverse field, not the datasheet's "max operating temp" line, which assumes a favorable permeance coefficient

keep Pc ≥ 2 where possible — thin magnets in open circuits (Pc < 1) sit near the knee even at room temperature

match coating to environment — Ni-Cu-Ni for indoor/dry, epoxy or parylene layers for humidity, SmCo for autoclave/marine/chemical duty

protect edges — chamfers on the magnet, lead-ins on the pocket, assembly fixtures that prevent slam-together

never load magnets structurally — carry loads through the housing; the magnet should only be held, never squeezed or bent

incoming inspection — flux measurement + visual coating check per lot catches marginal material before it's in 10,000 assemblies

had a field failure?

Send us the symptoms and the application details. We'll help you identify the root cause and spec a grade, coating, and geometry that won't repeat it — backed by full material and plating certification on every lot.

talk to an engineer →