Dysprosium- and Terbium-Free NdFeB: How Grain-Boundary Engineering Keeps Magnets From Demagnetizing at High Temperature While Slashing Rare-Earth Cost

Subtitle: Customers worry "without Dy/Tb it will demagnetize when hot" — first separate grain-boundary diffusion from heavy-rare-earth-free, then prove the reliability

1. First, Correct a Common Misconception: GBD and "Heavy-Rare-Earth-Free" Are Two Different Things

When many overseas customers hear "no dysprosium (Dy), no terbium (Tb)," the first reaction is: will it demagnetize at high temperature? And they often conflate "grain-boundary diffusion" with "heavy-rare-earth-free." This must be clarified first, because they are two completely different technology routes.

Grain Boundary Diffusion (GBD): Still uses Dy/Tb, but instead of melting Dy/Tb into the bulk magnet alloy, a Dy/Tb-containing coating or foil is placed on the magnet surface, and high temperature lets Dy/Tb diffuse inward along the grain boundaries, enriching only the grain-shell layer. The result: Dy/Tb usage drops sharply (typically 60–80% less), but it isn't fully removed — it is "low heavy rare earth," not "heavy-rare-earth-free."

Heavy-Rare-Earth-Free (Dy/Tb-free): No dysprosium, no terbium at all. It relies on the magnet's own microstructure design — grain refinement, grain-boundary phase engineering, composition optimization — to raise coercivity (resistance to demagnetization).

This article focuses on the latter, because that is exactly what customers worry about: "without Dy/Tb, is it reliable?"

2. The Traditional Approach: Using Dy/Tb for Heat, at a High Cost

To understand why Dy/Tb-free is viable, first see the cost of the traditional approach.

NdFeB's biggest weakness: its coercivity Hcj falls as temperature rises, with a temperature coefficient of roughly -0.55% to -0.65%/°C. As temperature climbs, Hcj drops, and the magnet can be pushed past its "knee point" by a reverse field and suffer irreversible demagnetization.

The traditional fix: dope Dy/Tb into the Nd2Fe14B main phase (forming Nd2-xDyxFe14B), using Dy/Tb's higher magnetocrystalline anisotropy field to push Hcj up and survive high temperature.

But the cost is twofold:

Cost one: expensive and scarce. Dy and Tb cost several to over ten times metal Nd, with highly concentrated supply and high geopolitical risk. Every bit used raises cost and amplifies supply-chain risk.

Cost two: sacrificed magnetic performance. Dy/Tb have lower magnetic moment than Nd; the more you add, the more remanence Br and maximum energy product (BH)max drop. Empirically, adding about 10 wt% Dy lowers Br by roughly 0.10–0.15 T while raising Hcj by about 5–8 kOe. It's trading magnetic performance for temperature stability — a poor bargain, especially for motor designs chasing high energy product.

3. The Physical Nature of Demagnetization: Not That the Magnet Is "Weak," But That the Operating Point Crosses the Knee

To dissolve customer worry, the most important thing to explain is: how does demagnetization actually happen?

Every NdFeB magnet has a demagnetization curve, and at its end is a "knee point." The magnet has an operating point in use (set by the magnetic-circuit design). When a reverse field — from self-demagnetization, external disturbance, and the Hcj drop caused by temperature rise, combined — pushes the operating point past the knee, irreversible demagnetization occurs: the magnet's force drops permanently and never returns.

Key insight: Resisting demagnetization depends on Hcj and the safety margin of the operating point, not on Br, and not on whether Dy/Tb was added. So the question "will it demagnetize without Dy/Tb" is itself imprecise — what should really be asked is: can the Dy/Tb-free magnet maintain sufficiently high Hcj and a sufficiently large operating-point margin?

The answer comes from the three microstructure routes below.

4. Dy/Tb-Free Route One: Grain Refinement

NdFeB coercivity is nucleation-controlled — reversed domains nucleate mainly at grain surfaces and grain-boundary defects, not by pinning inside the bulk. This gives grain refinement a lever.

The finer the grains, the more easily each grain is "isolated" by the grain-boundary phase; reversed domains can't propagate across grains, the nucleation field rises, and Hcj goes up. Reducing average grain size from about 10 μm to 3–5 μm typically raises Hcj by 1.5–2×. This isn't trading performance for elements — it's making the structure finer, with Br essentially intact.

The means are mainly powder treatment and sintering control: hydrogen decrepitation (HD) for crushing, jet milling (JM) for refinement, and precise control of sintering temperature and soak to avoid abnormal grain growth.

5. Dy/Tb-Free Route Two: Grain-Boundary Engineering (the Core)

This is the true core of the Dy/Tb-free route.

After sintering, NdFeB grains are separated by a rare-earth-rich (RE-rich) grain-boundary phase. If this layer is thin, continuous, and weakly magnetic, it isolates the main-phase grains — a reversed domain nucleated in one grain can't cross the boundary, and the whole magnet's Hcj rises. Conversely, if the boundary phase is discontinuous with direct main-phase connections, reversed domains travel freely and Hcj collapses.

Optimizing this boundary phase relies on three means:

Element control. Adding Cu, Al, Ga, etc., improves the wettability and continuity of the boundary phase so it coats grain surfaces more uniformly. Adding Nb, Zr, Ti, V, W, Mo, etc., forms borides or carbides that refine and pin the boundaries.

Dual-alloy method. The "main alloy" (providing the Nd2Fe14B main phase) and the "boundary alloy" (Nd-rich, containing Cu, Al, Ga and other boundary elements) are batched separately, blended as powder, then sintered. This lets the main phase and boundary phase be designed independently — main phase preserves energy product, boundary phase preserves Hcj, both achieved.

Process control. The powder-size distribution from HD+JM, oxygen control, and sintering/aging schedule all directly determine the boundary-phase morphology. Dy/Tb-free grades are more sensitive to the process window than traditional grades — which is exactly where the technical threshold lies.

6. Dy/Tb-Free Route Three: Composition Design to Control the Temperature Coefficient

Grain refinement and grain-boundary engineering raise room-temperature Hcj, but NdFeB's Hcj temperature coefficient is still negative — it still drops when hot. To make it "not drop too much" at high temperature, a third lever is needed: lowering the temperature coefficient itself.

The most effective means is cobalt (Co) substitution: replacing part of the Fe in Nd2Fe14-xCoxB with Co. Co significantly raises the Curie temperature Tc and simultaneously improves the negative temperature coefficients of both Br and Hcj. Combined with grain refinement and grain-boundary engineering, Dy/Tb-free magnets can stably operate at 150–180°C (equivalent to UH, EH temperature grades) without relying on Dy/Tb at all.

This is precisely the key that makes "Dy/Tb-free high-temperature grades" possible.

7. The Truth About Temperature Stability: High Room-Temp Hcj + Design Margin + Low Temperature Coefficient

Stringing the three routes together, a Dy/Tb-free magnet guarantees "no demagnetization when hot" through a three-step closed loop:

Step one, high room-temperature Hcj. Grain refinement + grain-boundary engineering push room-temperature Hcj to 20–30 kOe (1600–2400 kA/m), a notch above ordinary grades.

Step two, ample operating-point margin. In the magnetic-circuit design, keep the operating point — after load, temperature rise, and reverse field are combined — firmly above the knee. This is accomplished jointly by the motor designer and the magnet factory, not by a single magnet-side metric.

Step three, low temperature coefficient. Co slows Hcj's high-temperature decay. Example: if room-temperature Hcj is 30 kOe and the coefficient is -0.6%/°C, at 180°C Hcj is still about 12 kOe; with a reasonable load operating point, the margin remains ample.

In other words: a Dy/Tb-free magnet doesn't "force itself not to demagnetize" — it raises room-temperature Hcj, lowers the temperature coefficient, and leaves ample design margin, so it simply never touches the knee within the working temperature range.

8. How Reliability Is Proven — Responding to the Customer's Biggest Worry

When a customer asks "is it reliable," you can't just explain theory; you need evidence. A professional magnet factory proves a Dy/Tb-free grade is reliable through a quantifiable verification set:

Demagnetization-curve measurement. Provide the complete second-, third-, fourth-quadrant demagnetization curve and mark the knee position — the most intuitive evidence.

Thermal aging test. Hold at the target working temperature (e.g., 150°C) for 500–1000 hours, then measure Hcj and Br decay, typically requiring decay under 5%.

Open-circuit flux-loss test. Simulate the magnet's actual open-circuit state inside the motor and measure flux-loss rate at high temperature — closer to real duty than looking at Hcj alone.

Batch consistency + third-party certification. Keep samples and traceability per batch; key grades get third-party testing. Reliability isn't single-piece compliance but million-piece-level consistency.

Collaborative design. The most responsible approach is for the factory to help the customer calculate the load operating point versus knee relationship and provide a usable temperature-load envelope, rather than just quoting an Hcj number.

9. The Cost Picture: What's Cut Is the Most Expensive and Scarce Part

Why is Dy/Tb-free worth doing? Because what's cut is the most expensive, scarcest, highest-geopolitical-risk part of raw-material cost.

Dy/Tb cost several to over ten times metal Nd. Removing them significantly lowers magnet raw-material cost while shifting the supply chain from "at the mercy of others" to "ordinary NdFeB suffices." For high-volume robotics, motors, and automotive applications, this is real cost reduction.

But be honest: Dy/Tb-free isn't "zero cost" — grain refinement and grain-boundary engineering raise process cost (process trades for raw material). Overall, total cost is still clearly lower than Dy/Tb-containing grades, and the more volume, the more it saves.

The honest boundary: when working temperature exceeds about 180–200°C, or when extremely high Hcj is required (some extreme servos, aerospace), a small amount of Dy/Tb, or the GBD (grain-boundary diffusion) route for a "low heavy rare earth" compromise, may still be needed. Technology selection must follow duty; no single solution fits all.

10. Selection Guidance and Conclusion

Practical selection logic for customers:

· Working temperature ≤120°C: Ordinary Dy/Tb-free high-Hcj grade (N, M grade plus grain refinement) suffices.

· Working temperature 120–150°C: Dy/Tb-free + Co high-temperature grade (corresponding to SH, UH grades), no Dy/Tb needed.

· Working temperature 150–180°C: Dy/Tb-free + Co-strengthened grade (corresponding to EH grade), requiring strict verification of operating-point margin.

· Working temperature >180°C or ultra-high Hcj demand: Evaluate a small amount of Dy/Tb or the GBD low-heavy-rare-earth route.

Dy/Tb-free NdFeB is not "cutting corners" — it shifts reliability from "adding expensive heavy metal" to "microstructure design." Same high-temperature non-demagnetization, but raw-material cost drops sharply. The reliability concern customers worry about is essentially an engineering problem answerable with data, curves, and test loops — not mysticism.

For Dy/Tb-free grade selection, demagnetization-curve evaluation, or Co-strengthened high-temperature magnet solutions, please feel free to reach out — FAIZEAL can provide technical support.