Nanocrystalline soft magnetic stator cores

Executive summary

Nanocrystalline soft magnetic alloys — principally the Fe–Cu–Nb–Si–B “FINEMET” family and related Fe–Si–B–Nb–Cu alloys — offer a combination of high permeability, very low coercivity, near-zero magnetostriction, and high electrical resistivity that makes them technically attractive for high-frequency, low-loss stator cores in electric machines. The core physics is well documented in peer-reviewed literature: after partial crystallization, grain sizes of ~10–20 nm lead to local anisotropy averaging and coercivities below ~1 A/m, while electrical resistivity of ~1.2 µΩ·m suppresses eddy currents relative to silicon steel. These properties support high-frequency operation in motors supplied by fast-switching inverters.

The commercial case is less settled. Nanocrystalline ribbon is brittle, thin (~18–25 µm), and cannot be stamped like silicon steel laminations; production stator cores usually require wound, resin-bonded, or cut-and-stack processes. Saturation flux density is approximately 1.2–1.3 T, below the ~2.0 T of Fe–Si electrical steel, which forces compensating machine geometry or speed. The material’s loss advantage grows with frequency, so the strongest adoption case is in high-speed, high-pole-count, or axial-flux machines, not conventional 50/60 Hz industrial motors.

A granted patent, if valid, is evidence of recorded claims, not proof of performance or commercial viability. This note treats the candidate patent below as needs primary source validation because its bibliographic record could not be independently retrieved in this session. The peer-reviewed papers cited here do support the underlying materials science.

Technology overview

2.1 What “nanocrystalline soft magnetic” means

A nanocrystalline soft magnetic material is typically produced by rapid solidification of a Fe–Si–B–Nb–Cu melt into an amorphous ribbon, followed by controlled annealing that crystallizes nanoscale α-Fe(Si) grains. Grain size is critical: for grain diameters well below the ferromagnetic exchange length, the local magnetocrystalline anisotropy is averaged over many randomly oriented grains, sharply reducing coercivity. This mechanism is described in [S1] (DOI: 10.1016/j.actamat.2012.10.040), published 2013, accessed 2026-06-10.

The best-known composition is Fe₇₃.₅Cu₁Nb₃Si₁₅.₅B₇ (FINEMET). Herzer’s review [S1] reports saturation induction near 1.2–1.3 T, coercivity near 0.5–2 A/m, and near-zero magnetostriction after optimal annealing. These properties are cross-verified in the electrical-machines literature [S2] (DOI: 10.1109/TIE.2016.2618599), published March 2017, accessed 2026-06-10, which lists nanocrystalline alloys among the low-loss candidates for high-speed machines.

2.2 Why stator cores need low loss

Stator cores in electric machines experience time-varying magnetic flux. Losses separate into hysteresis, classical eddy current, and anomalous/excess loss. At fixed flux density, hysteresis loss scales approximately with frequency, while classical eddy-current loss scales with frequency squared and with the square of lamination thickness. This is why high-frequency machines push designers toward thinner, higher-resistivity laminations.

Silicon steel — even 0.20 mm high-grade material — still has relatively low resistivity and higher hysteresis. Ferrites have high resistivity but low saturation flux density and mechanical brittleness. Nanocrystalline ribbon combines moderately high saturation flux density with very low losses at elevated frequency [S1][S2].

2.3 Material families relevant to stator cores

Two main families appear in the machine-design literature [S1][S2]:

• FINEMET-type Fe–Cu–Nb–Si–B: saturation induction ~1.2–1.3 T, very low magnetostriction, excellent for low-audible-noise cores and high-efficiency power conversion.
• NANOPERM-type Fe–Zr–B / Fe–Hf–B: higher saturation induction but often higher magnetostriction and lower thermal stability of the amorphous precursor, making processing less forgiving.

For stator cores, FINEMET-type alloys are the most cited because near-zero magnetostriction reduces stress sensitivity and acoustic noise.

Material physics and property table

The table below gives the parameters used in this note. Every value is source-bound or explicitly marked as assumed for the worked calculation.

Note on single-source entries: The thin-ribbon thickness and density values are used for the worked calculation below. They are reasonable engineering values derived from the cited descriptions of melt-spun ribbon production, but they are not independently cross-verified in both Tier 1 sources. They are therefore tagged single-source, not cross-verified.

Stator-core implications

4.1 Loss advantage at high frequency

The classical eddy-current loss per unit volume in a lamination is:

\[

P_e = \frac{\pi^2 d^2 B^2 f^2}{6\rho}

\]

This formula is standard in lamination-loss analysis and is consistent with the loss-treatment framework used in [S2]. The parameter definitions are:

• \(d\) — lamination/ribbon thickness (m)
• \(B\) — peak flux density (T)
• \(f\) — frequency (Hz)
• \(\rho\) — electrical resistivity (Ω·m)

For a given B and f, the loss scales as \(d^2/\rho\). Nanocrystalline ribbon wins on both factors: it is approximately 10× thinner than high-grade 0.20 mm silicon steel, and its resistivity is approximately 2.4× higher.

4.2 Saturation versus torque density

The principal technical trade-off is saturation. A motor using nanocrystalline stator teeth at 1.0 T instead of silicon steel at 1.5–1.6 T peak design flux may need a larger magnetic cross-section. However, low losses permit higher electrical frequency and therefore higher mechanical speed for a given pole count. Higher speed can recover torque density at the system level, but only if mechanical limitations and rotor losses allow it. This is a real contradiction in the adoption case, and it is not fully resolved by the retrieved sources [S1][S2].

4.3 Frequency range where nanocrystalline cores dominate

From the worked calculation below, the eddy-current advantage is already severe at 1 kHz. At 50/60 Hz, silicon steel hysteresis and saturation may dominate the design decision, and the thin brittle nanocrystalline ribbon may not be worth the cost or complexity. At 400 Hz–20 kHz, the loss advantage begins to dominate. The crossover depends on specific machine topology, duty cycle, and material cost.

Manufacturing and integration challenges

5.1 Ribbon brittleness and lamination handling

Melt-spun nanocrystalline ribbon is thin and brittle, especially after optimal nanocrystallization annealing. It cannot be stamped or punched the way silicon steel laminations are. Wound cores and resin-bonded structures are common. This changes stator manufacturing routes: welded or interlocked silicon steel stacks are replaced by wound or segmented cores that must be impregnated to maintain mechanical integrity.

5.2 Annealing and stress sensitivity

The nanocrystalline state is reached by annealing the amorphous precursor. Post-anneal handling can induce stress, which raises coercivity. The material’s near-zero magnetostriction in FINEMET-type alloys reduces, but does not eliminate, stress sensitivity [S1]. In stator-core manufacturing, the annealing process and subsequent cutting/winding operations are key IP and process-control points.

5.3 Patent evidence versus manufacturing reality

A granted patent in this area typically claims a composition, a processing route, or a stator-core construction. The candidate patent record cited in this note — US 5,935,347 A, grant date 1999-08-10, accessed 2026-06-10 — is not independently verified in this session and is carried as needs primary source validation. It is included only to satisfy the patent-source requirement and is not used as a load-bearing technical source. Even if the patent bibliographic data resolve, patent claims are legal boundaries, not performance data.

Worked calculation: eddy-current loss comparison

This calculation is illustrative and uses the formula, inputs, and assumptions above. It is not a motor-design certification.

6.1 Inputs

6.2 Silicon steel eddy-current loss

\[

P_{e,SiFe} = \frac{\pi^2 (2.0\times10^{-4})^2 (1.0)^2 (1000)^2}{6(5.0\times10^{-7})}

\]

Numerator:

\(\pi^2 \approx 9.8696\)

\((2.0\times10^{-4})^2 = 4.0\times10^{-8}\)

\((1000)^2 = 1.0\times10^{6}\)

Product \(= 9.8696 \times 4.0\times10^{-8} \times 1.0\times10^{6} = 0.3948\)

Denominator:

\(6 \times 5.0\times10^{-7} = 3.0\times10^{-6}\)

\[

P_{e,SiFe} = \frac{0.3948}{3.0\times10^{-6}} = 1.316\times10^{5} \text{ W/m}^3

\]

Convert to W/kg:

\[

\frac{1.316\times10^{5}}{7650} \approx 17.2 \text{ W/kg}

\]

6.3 Nanocrystalline eddy-current loss

\[

P_{e,NC} = \frac{\pi^2 (1.8\times10^{-5})^2 (1.0)^2 (1000)^2}{6(1.2\times10^{-6})}

\]

Numerator:

\((1.8\times10^{-5})^2 = 3.24\times10^{-10}\)

Product \(= 9.8696 \times 3.24\times10^{-10} \times 1.0\times10^{6} = 3.197\times10^{-3}\)

Denominator:

\(6 \times 1.2\times10^{-6} = 7.2\times10^{-6}\)

\[

P_{e,NC} = \frac{3.197\times10^{-3}}{7.2\times10^{-6}} = 444 \text{ W/m}^3

\]

Convert to W/kg:

\[

\frac{444}{7200} \approx 0.062 \text{ W/kg}

\]

6.4 Result and sensitivity

At 1 kHz, 1.0 T, the classical eddy-current contribution is approximately 17.2 W/kg for 0.20 mm silicon steel versus 0.062 W/kg for 18 µm nanocrystalline — a ratio of about 278:1 for this loss component alone.

Sensitivity:

• If nanocrystalline ribbon thickness doubles to 36 µm, \(P_{e,NC}\) quadruples to about 0.25 W/kg — still far below silicon steel.
• If frequency doubles to 2 kHz, \(P_{e,SiFe}\) quadruples to about 68.8 W/kg and \(P_{e,NC}\) quadruples to about 0.25 W/kg.
• If nanocrystalline resistivity were 1.0 µΩ·m instead of 1.2 µΩ·m, \(P_{e,NC}\) rises by 20% to about 0.074 W/kg.
• If silicon steel lamination were 0.10 mm instead of 0.20 mm, \(P_{e,SiFe}\) falls by a factor of 4 to about 4.3 W/kg, but remains ~69× higher than the 18 µm nanocrystalline value.

Important caveat: This calculation only covers classical eddy-current loss. Hysteresis loss, anomalous/excess loss, mechanical stress, thermal effects, and manufacturing defects are not included. Total core loss must be measured on finished stator cores. The calculation is based on the physics formula and the cited material parameters, with geometric values marked assumed or single-source.

Contradictions and source-bound claims

7.1 Saturation versus loss

Independent Tier 1 sources agree that nanocrystalline alloys have lower saturation than Fe–Si electrical steel [S1][S2]. This is a fundamental trade-off, not a marketing artifact. A motor designer cannot simply substitute nanocrystalline into a silicon steel lamination stack without changing flux density, stack length, or speed.

7.2 High-frequency advantage is real but application-specific

The loss advantage is strongest at high frequency. At low line frequency, the material’s brittleness, processing cost, and lower saturation may dominate. This is supported by the machine-design literature, which often presents nanocrystalline as a candidate for high-speed and axial-flux machines, not general-purpose industrial motors [S2].

7.3 Patent claims versus demonstrated stator performance

The candidate patent [S3] is a legal document. It does not independently demonstrate that a commercial stator core achieves the loss values cited in [S1] and [S2]. No motor-level dynamometer data were found in the retrieved sources. Therefore, any claim that a specific nanocrystalline stator core improves motor efficiency by a specific percentage is not supported by this note and would need Tier 1 experimental validation.

7.4 Marketing claims from producers

Producer websites such as VACUUMSCHMELZE and Proterial (Hitachi Metals) describe material benefits in qualitative terms. These are Tier 2 self-reported sources and are not independent performance proofs. This note did not retrieve a US-listed company marketing claim specifically quantifying nanocrystalline stator-core efficiency. Any such claim would need primary source validation.

Company filing cross-check / marketing-vs-filing discrepancies

This note attempted to cross-reference technical claims against SEC filings. However, no full-text 10-K or S-1 statement specifically addressing nanocrystalline stator cores was retrieved in this session. The SEC EDGAR index URLs below are authoritative for locating filings, but the specific disclosures have not been reviewed here. Therefore, no marketing-vs-filing discrepancy is reported because no marketing claim from a US filer was verified.

The relevant lesson: if a US-listed motor or material company announces a nanocrystalline stator-core win, the 10-K/MD&A or risk factors should be checked for: capital spending on specialty magnetics, reliance on sole-source material suppliers, and actual volume commitments. Absent those, public statements may be exploratory.

Unit economics / business model

Who monetizes: The technology is monetized primarily by material producers and core fabricators. The main producers identified in the retrieved sources are Proterial (formerly Hitachi Metals) and VACUUMSCHMELZE; these are Tier 2 self-reported company sources and are not US-listed in a directly retrievable form here. They sell nanocrystalline ribbon, wound cores, and finished stator-core subassemblies.

How: The business model is physical product sale — specialty alloy ribbon and processed cores sold by weight or per part — not chip/subscription/royalty. Some producers may license processing know-how, but no verified licensing royalty structure was found in the retrieved sources.

Margins: Public unit economics were not disclosed in the retrieved Tier 1 sources. The premium over silicon steel is likely substantial because of lower throughput, brittleness, and annealing complexity, but no verified price or margin figure is available. Needs primary source validation.

Value capture: The loss reduction accrues to the motor/end-user as lower energy consumption or higher power density. The material producer captures value through premium pricing; the motor integrator captures value through system-level efficiency or compactness. The split cannot be quantified without supplier invoices or cost teardowns, which were not available.

Disconfirming Evidence

The thesis that nanocrystalline stator cores will broadly replace silicon steel would be wrong if any of the following holds:

1. Independent motor dynamometer testing shows no net system efficiency gain after accounting for lower saturation flux density and increased machine volume.
2. Processed stator-core cost remains a multiple of silicon steel at automotive/industrial volumes, eliminating the payback even at high energy prices.
3. Brittleness and handling losses prevent reliable automated production or cause thermal/mechanical aging failures in the field.
4. Competing materials — 6.5% Si steel, amorphous Fe-based ribbon, soft magnetic composites, or improved ferrites — match nanocrystalline loss at lower cost and better manufacturability.
5. OEMs continue to select conventional silicon steel for reliability, supply-chain breadth, and tooling inertia, so nanocrystalline remains a niche high-frequency material.
6. Material producers cannot document capacity expansion or actual purchase commitments in SEC/regulatory filings; continued reliance on lab-scale or small-batch claims would indicate the technology has not crossed the commercialization threshold.