NdFeB Magnet Demagnetisation Failure: Temperature Coefficient, Coercivity and Corrosion Root Cause

Overview #

The failure mode that most procurement teams misattribute when NdFeB magnets underperform in service is thermal demagnetisation — but in the majority of cases we investigate, the root cause is not temperature exceedance. It is coercivity grade mismatch at the specification stage, compounded by corrosion-driven flux loss that develops over 6–18 months in humid or chemically aggressive environments. Buyers who specify only remanence (Br) and maximum energy product (BHmax) are leaving out the two parameters that actually determine whether a magnet survives its operating environment: intrinsic coercivity (Hcj) and corrosion protection class. Getting those two right at the sourcing stage eliminates the majority of field failures we see from Chinese-sourced NdFeB.

Failure Mode 1: Thermal Demagnetisation from Coercivity Grade Mismatch #

The most common thermal failure in NdFeB magnets is not caused by exceeding the Curie temperature (typically 310–340°C for standard grades). It is caused by operating above the magnet’s maximum working temperature — a threshold that is grade-specific and directly controlled by intrinsic coercivity (Hcj). For a standard N35 grade, maximum working temperature is 80°C. For N35H, it is 120°C. For N35SH, 150°C. For N35UH, 180°C. For N35EH, 200°C. These are not conservative estimates — they are the temperatures at which irreversible flux loss begins under a typical permeance coefficient of 1.0–2.0.

The temperature coefficient of remanence (α) for sintered NdFeB is approximately −0.11% to −0.13% per °C, and the temperature coefficient of coercivity (β) is approximately −0.55% to −0.65% per °C. The coercivity coefficient is five times more sensitive to temperature than remanence. This means a magnet operating at 20°C above its rated maximum loses coercivity far faster than it loses flux density — and once coercivity drops below the demagnetising field present in the circuit, the loss is irreversible.

In our supplier qualification program, we test for irreversible flux loss by cycling magnets through their rated maximum working temperature for 100 hours, then measuring flux at 20°C before and after. The acceptance threshold we apply is ≤3% irreversible flux loss after this cycle. Suppliers who cannot provide this data — not just a datasheet claim, but actual lot-specific test results — are not qualified for motor or sensor applications.

The IEC Standards framework, specifically IEC 60404-8-1, covers magnetic alloy and steel specifications relevant to permanent magnet characterisation. Buyers sourcing for automotive or industrial motor applications should also reference SAE International standards for magnet performance in traction environments.

Most procurement teams over-specify BHmax and under-specify Hcj. The grade designation tells you the energy product. It does not tell you the coercivity class — that requires reading the full B-H demagnetisation curve, not just the headline number on the datasheet.

For applications involving pump-valve-seals or magnetic coupling assemblies where operating temperature fluctuates with process fluid temperature, the SH or UH grade is almost always the correct specification — even when the nominal operating temperature appears to fall within N35H range. Thermal spikes during startup and shutdown cycles are the trigger for irreversible loss, not steady-state temperature.

Failure Mode 2: Corrosion-Driven Flux Loss and Coating Specification Errors #

NdFeB is the most corrosion-susceptible of the commercial permanent magnet families. The neodymium-rich grain boundary phase oxidises preferentially in the presence of moisture, and once oxidation penetrates the grain boundary network, the magnet physically disintegrates — not gradually, but in flakes and powder. We have seen magnets in IP54-rated enclosures fail within 14 months in coastal industrial environments where relative humidity consistently exceeded 85%.

The standard coating options and their protection levels are not equivalent, and this is where most sourcing decisions go wrong. Nickel-copper-nickel (Ni-Cu-Ni) triple-layer electroplating is the most common coating on Chinese-sourced NdFeB and provides adequate protection for indoor, low-humidity environments. It does not provide adequate protection for outdoor, marine, or chemically aggressive environments. Epoxy coating provides better chemical resistance but lower mechanical durability. Zinc plating is cheaper and significantly less protective — we see it specified on cost-optimised batches where the buyer did not explicitly require Ni-Cu-Ni, and the supplier defaulted to the lower-cost option.

The salt spray test per ASTM International ASTM B117 is the standard acceptance test for coating integrity. For Ni-Cu-Ni coated NdFeB, the minimum acceptable result for general industrial use is 48 hours without red rust. For marine or outdoor applications, we require ≥96 hours. Suppliers who quote “salt spray tested” without specifying hours and rust criteria are providing meaningless documentation.

In our qualification program, we have seen suppliers pass initial sample approval with 72-hour salt spray results and then deliver production batches where coating thickness had dropped from the specified 15–20 µm to 8–10 µm due to a plating line process change. The COA showed no change because coating thickness was not a line item on the standard COA format. The failure appeared in the field 11 months later as surface oxidation and flux loss of approximately 8–12% — well above the ≤3% acceptance threshold. The root cause was not material grade — it was an undocumented process change at the plating subcontractor, which the magnet supplier did not disclose because they did not consider it a specification-relevant change.

This is the sourcing friction that standard incoming inspection does not catch. Coating thickness measurement (eddy current method, per ISO Standards ISO 2360) should be a mandatory incoming inspection item for any NdFeB application in a humid or chemically aggressive environment. The measurement takes less than two minutes per sample and catches the most common production-volume substitution we see from Chinese suppliers.

Most Western buyers do not realise that SAC China Standards GB/T 13298 and GB/T 5591.3 govern NdFeB magnet specifications in China, and that the dimensional and magnetic tolerances in these standards are not identical to IEC or ASTM equivalents. A supplier declaring GB/T compliance is not automatically declaring IEC compliance. The gap is usually small for magnetic properties but can be significant for dimensional tolerances — particularly for thin disc magnets where ±0.1 mm tolerance under GB/T may not meet a ±0.05 mm engineering drawing requirement.

Failure Mode 3: Mechanical Fracture and Handling-Induced Microcracking #

Sintered NdFeB has a flexural strength of approximately 200–250 MPa and a fracture toughness of 1.0–2.0 MPa·m^0.5 — comparable to a hard ceramic, not a metal. It chips, cracks, and fractures under impact loads that would be trivial for steel. This is not a defect. It is an intrinsic material property. The failure mode we see in production is not bulk fracture — it is microcracking during assembly, which is invisible at incoming inspection but propagates under thermal cycling or vibration to produce flux loss and eventual mechanical failure.

The detection method for microcracking is not visual inspection. It is flux mapping — measuring the spatial distribution of flux density across the magnet face using a Hall probe array or fluxmeter. A microcracked magnet will show localised flux anomalies of 5–15% below the mean value at the crack location. We require flux mapping on 100% of magnets for motor and sensor applications, and on a 5% AQL sample for general industrial use.

Buyers sourcing NdFeB for sensors-detection applications — particularly position sensors and encoders where field uniformity is critical — should specify flux uniformity tolerance explicitly: typically ±2–3% across the active face. This is not a standard line item on a Chinese supplier’s COA and will not be tested unless explicitly contracted.

Production Failure Scenario: Motor Magnet Demagnetisation in HVAC Compressor Application #

A European HVAC equipment manufacturer sourced N35H grade NdFeB arc magnets from a Chinese supplier for a brushless DC compressor motor. The specification called for Br ≥ 1.17 T, Hcj ≥ 1353 kA/m, and Ni-Cu-Ni coating. Initial sample approval passed all parameters. Production volume delivery began at month 3.

At month 9, field returns began showing motor efficiency loss of 12–18% in units installed in rooftop applications in Southeast Asia. Teardown analysis showed irreversible flux loss of 9–14% in the arc magnets, with visible surface oxidation at the magnet edges despite intact coating on the flat faces.

Root cause analysis identified two concurrent failures. First, the supplier had substituted raw material from a different NdFeB powder compounder at month 5, which reduced actual Hcj from the specified ≥1353 kA/m to approximately 1100–1150 kA/m — still within the N35H designation range at the lower boundary, but insufficient for the actual demagnetising field present in the motor circuit at peak load current. Second, coating thickness at the arc edges (the highest-stress geometry for plating) had dropped to 6–8 µm versus the specified 15 µm minimum, allowing moisture ingress at the edge chamfer.

The buyer’s incoming inspection had tested Br and BHmax but not Hcj directly — because Hcj measurement requires a full B-H curve tracer, which most incoming inspection labs do not have. The coating thickness had not been specified as an incoming inspection item.

Corrective action required: (1) Hcj measurement added to incoming inspection using a permeameter per IEC Standards IEC 60404-14, with a minimum acceptance threshold of 1320 kA/m (not the datasheet minimum of 1353 kA/m, to provide margin for measurement uncertainty); (2) coating thickness specified as 15 µm minimum with ±2 µm tolerance, measured at edge locations, not face centre; (3) three consecutive batch COAs with Hcj data required before re-qualification of the supplier.

The total cost of the field failure — returns, teardown, re-engineering, and re-qualification — exceeded the cost savings from the Chinese source by a factor of approximately 4:1 over the affected production run.

Practical Guidance for Buyers #

When sourcing NdFeB magnets from China, the first specification to request from suppliers is not the BHmax grade — it is the full demagnetisation curve (B-H curve) with Hcj data for the specific lot, not just the grade datasheet. Most buyers ask for the datasheet. The datasheet tells you the grade minimum. It does not tell you where the actual lot sits within that range, and for high-coercivity grades (SH, UH, EH), lot-to-lot Hcj variation of 5–8% is common.

The sourcing mistake with the most consistent real-world consequence is accepting coating specification by grade name (“Ni-Cu-Ni”) without specifying minimum thickness and test method. As the compressor motor case above demonstrates, a coating that passes visual inspection and even a 48-hour salt spray test at 15 µm can fail at 6–8 µm in 9 months of humid service. Specify coating thickness at edge locations, not face centre — that is where plating is thinnest and where corrosion initiates.

Before committing to volume order, require: (1) lot-specific B-H curve with Hcj value, (2) salt spray test report per ASTM B117 with hours and rust criteria stated, (3) coating thickness measurement report with edge-location data, and (4) three consecutive batch COAs showing Hcj consistency. Suppliers who cannot provide items 1 and 3 are not qualified for any application above 80°C or in humidity above 60% RH.

Frequently Asked Questions #

Q1: What is the most important specification to verify on a NdFeB magnet COA from a Chinese supplier?

A: Intrinsic coercivity (Hcj), not BHmax. BHmax is easier to verify with basic equipment and easier to manipulate by grade-labelling. Hcj determines whether the magnet survives its operating temperature and demagnetising field — and it requires a full B-H curve tracer to measure properly.

Q2: How do I select the correct NdFeB grade for an application with a maximum operating temperature of 130°C?

A: N35H is rated to 120°C and N35SH to 150°C — so 130°C falls in a gap where N35H is marginal and N35SH is the conservative choice. Per the comparison table above, N35SH carries a minimum Hcj of ≥1592 kA/m versus N35H at ≥1353 kA/m. For any application with thermal spikes above nominal, specify SH grade. The cost premium over H grade from Chinese suppliers is typically 8–12% — far less than the cost of a field failure.

Q3: What is the most common quality failure in Chinese-sourced NdFeB at production volume?

A: Coating thickness reduction at the plating subcontractor level, which does not appear on a standard COA. We have seen this cause flux loss of 8–12% in field service within 9–14 months. The fix is simple: specify coating thickness as a COA line item with edge-location measurement, and spot-check incoming batches with an eddy current gauge per ISO Standards ISO 2360.

Q4: What test documentation should I require before approving a Chinese NdFeB supplier for a motor application?

A: Lot-specific B-H curve with Hcj value, salt spray test report per ASTM International ASTM B117 stating hours and rust criteria (minimum 48h for indoor, 96h for outdoor/marine), coating thickness report with edge-location data, and flux mapping results if field uniformity is critical. Three consecutive batch COAs showing Hcj consistency across lots is the minimum for qualification.

Q5: Does GB/T compliance mean a Chinese NdFeB magnet meets IEC or ASTM specifications?

A: No. SAC China Standards GB/T tolerances for NdFeB are not identical to IEC equivalents, particularly for dimensional tolerances on thin or complex geometries. Always cross-reference your engineering drawing tolerances against the specific GB/T standard the supplier is citing — do not assume equivalence.

Published by sinoraw.com Technical Team | Request a sourcing consultation