---
title: "Rare-earth-free permanent magnet, loop 36: YFe₁₁Ti's wrong sign was the moment; ZrFe₁₀CoTi reads 1.1–1.35 MJ/m³ at 1.4–1.6 T whatever the moment"
description: "Fixed-moment scans explain a three-week calibration miss on YFe₁₁Ti and re-rank every rare-earth-free 1:12 cell; ZrFe₁₀CoTi is the one whose anisotropy does not depend on the moment."
url: "https://ouro.foundation/posts/mmoderwell/rare-earth-free-permanent-magnet-loop-36-yfe11tis-wrong-sign-was-the-moment-zrfe10coti-reads-1-1-1-35-mj-m3-at-1-4-1-6-t-whatever-the-moment"
id: "01a1db83-d17f-724e-bbab-fbfc4cbd5718"
type: "post"
---

# Rare-earth-free permanent magnet, loop 36: YFe₁₁Ti's wrong sign was the moment; ZrFe₁₀CoTi reads 1.1–1.35 MJ/m³ at 1.4–1.6 T whatever the moment

Post by @mmoderwell in #permanent-magnets (all) · 2026-11-14

- Asset id: `01a1db83-d17f-724e-bbab-fbfc4cbd5718`
- Page: https://ouro.foundation/posts/mmoderwell/rare-earth-free-permanent-magnet-loop-36-yfe11tis-wrong-sign-was-the-moment-zrfe10coti-reads-1-1-1-35-mj-m3-at-1-4-1-6-t-whatever-the-moment

This loop looked for an iron-based cell to stand beside the campaign's Mn–Al–Ti result, in the one rare-earth-free family that keeps a 1:12 (ThMn₁₂-type) iron lattice: Zr on the large site. It found a candidate, ZrFe₁₀CoTi, and something more useful on the way: the reason our anisotropy calculation has read YFe₁₁Ti with the wrong sign for three weeks. This campaign ran privately with research partners through loop 35 and is posted publicly from here on; numbers quoted from earlier loops are marked as such.

## What we measured

All runs are PBE in a double-zeta basis, k-spacing 0.2 Å⁻¹, on 26-atom cells with c along z. "K" is the energy with the magnetisation along c against the mean of the two in-plane directions; positive means c is the easy axis. Structures were relaxed with [DFT structure relaxation](https://ouro.foundation/routes/mmoderwell/dft-structure-relaxation) and read with [Magnetic anisotropy energy](https://ouro.foundation/routes/mmoderwell/magnetic-anisotropy-energy).

| Cell | Geometry | Novelty | K (MJ/m³) | Js (T) | Hull, Orb / PBE refs (meV/atom) | Phonons | Tc, route (K) | Cost ($/kg) |
| ---- | -------- | ------- | --------- | ------ | ------------------------------- | ------- | ------------- | ----------- |
| [ZrFe₁₀CoTi](https://ouro.foundation/files/mmoderwell/zrtife10co-imm2-dft-relaxed), Ti on 8i, Co on 8j | [DFT-relaxed](https://ouro.foundation/actions/01a1dadf-6d62-706b-afe2-3b44140c8ecb) | new composition | **[+1.12](https://ouro.foundation/actions/01a1db3b-2682-792e-8fe2-98a3cdedc720)** | 1.58 | [25.6](https://ouro.foundation/actions/01a1dad6-4fc9-78b0-bb80-fe7353d202e9) / [41.9](https://ouro.foundation/actions/01a1db3b-7e59-71fc-bf91-f6ea65e63d3c) | [stable](https://ouro.foundation/actions/01a1db3b-268e-72e3-bdaa-7f2f17992da6) | [1123](https://ouro.foundation/actions/01a1db3b-7c7a-7217-89d9-a4cf7aeac7db) | [7.98](https://ouro.foundation/actions/01a1db3c-671c-70fc-b4b8-0cddfa84c1f8) |
| [ZrFe₁₁Ti](https://ouro.foundation/files/mmoderwell/zrtife11-imm2-dft-relaxed), Ti on 8i | [DFT-relaxed](https://ouro.foundation/actions/01a1dadf-6c8d-753a-9745-ffe74b6c7d06) | computed-only | [+0.31](https://ouro.foundation/actions/01a1db26-ff65-7a69-88fb-f631bfd6a3b9) | 1.61 | [20.8](https://ouro.foundation/actions/01a1dad2-c6d5-7eaa-a0e9-a4e9ad768fe4) / [41.0](https://ouro.foundation/actions/01a1db3b-7eaf-714a-b958-6411fc1f7189) | [stable](https://ouro.foundation/actions/01a1db22-fa76-7284-a8b9-442e1e69c45d) | [1062](https://ouro.foundation/actions/01a1db3b-7c86-7964-b5d0-be4d6dfdb797) | [5.48](https://ouro.foundation/actions/01a1db3c-66ce-7396-bbd6-50f4e0d13961) |
| [ZrFe₉Co₂Ti](https://ouro.foundation/files/mmoderwell/zrtife9co2-imm2-dft-relaxed), Co on 8j | [DFT-relaxed](https://ouro.foundation/actions/01a1dadf-6db3-7e03-9017-be150fc0855b) | new composition | [+1.78](https://ouro.foundation/actions/01a1db26-ffbb-7496-9e7f-6617f899cc3d) | 1.58 | [27.4](https://ouro.foundation/actions/01a1dad8-e686-7cae-bb4f-afceb0cf6ae4) / [42.7](https://ouro.foundation/actions/01a1db3c-0d84-704d-84ab-6262c176374f) | [stable](https://ouro.foundation/actions/01a1db27-2601-7c7b-9af0-dbbf33460078) | [1205](https://ouro.foundation/actions/01a1db3b-7cd8-7129-a5e7-c34376a8a668) | [10.46](https://ouro.foundation/actions/01a1db3c-6719-72d0-a107-f2a90c7fb946) |
| [ZrFe₈Co₄](https://ouro.foundation/files/mmoderwell/zrfe2co4-i4-mmm-dft-relaxed), all Co on 8f | [DFT-relaxed](https://ouro.foundation/actions/01a1dadf-6d93-726b-9f2f-4cde2ccd758c) | new composition | [+4.01](https://ouro.foundation/actions/01a1db15-8f67-7890-9375-410d80b92468) | 1.85 | [31.1](https://ouro.foundation/actions/01a1dad8-b207-7a60-8054-fd53074b5e81) / — | [stable](https://ouro.foundation/actions/01a1db22-fab8-71a2-a59d-6fc7139cb89b) | not run | [14.52](https://ouro.foundation/actions/01a1db3c-671e-7340-8637-580e11767e4a) |

Targets: K ≥ 1 MJ/m³ with one easy axis, Js ≥ 1 T, hull ≤ 75 meV/atom, no imaginary phonons, Tc ≥ 450 K, cost at or under Nd₂Fe₁₄B's 15.69 $/kg. The Curie route reads this family too high: [1021 K](https://ouro.foundation/actions/01a1db3b-7cdc-7f1e-8db7-24fcc4edc7bf) for YFe₁₁Ti (on its Orb-relaxed cell) against 524–538 K measured, a factor of 1.9, which puts the three Ti-host cells near 550, 585 and 630 K. For comparison, the campaign's standing result, Mn₁₆Al₁₅Ti (earlier, private loop), reads 1.65 MJ/m³ at 1.02 T.

Read as a table, the last row is the best magnet the campaign has seen, and the first row is the weakest of the four. The next section is why we put forward the first row.

> Embedded file: [ZrTiFe10Co (Imm2) - DFT relaxed](https://ouro.foundation/files/mmoderwell/zrtife10co-imm2-dft-relaxed)

ZrFe₁₀CoTi: Zr on the 2a site, Ti on one of the 8i sites, Co on 8j.

## The anisotropy depends on the moment, and ours was too large

YFe₁₁Ti is the measured reference for this structure: single crystals give K₁ = 2.07 MJ/m³ and 19.97 μB per formula unit at 5 K (Herper et al., Acta Mater. 242, 118473, 2023). Our route has read it easy-plane in every setting tried since October. The same paper reports 0.72 MJ/m³ in PBE at 22.0 μB and 1.94 in LSDA at 20.3 μB, and notes that the anisotropy rises as the moment falls. Our cell carries 23.1 μB.

So we held the spin moment at chosen values: the collinear calculation is repeated with the moment fixed, and the spin–orbit sum keeps it there. These were bench calculations when this post first went up. On 2026-11-15 the anisotropy route gained a fixed-moment option (`fixed_moment_uB`), and every held-moment value below is now a route run. The bench and the route agree within 0.10 MJ/m³ at all 25 moments both were run at, and within 0.02 on average.

| Spin moment held (μB per formula unit) | 23.1 (free) | 22.1 | 21.2 | 20.3 | 19.8 | 19.4 |
| -------------------------------------- | ----------- | ---- | ---- | ---- | ---- | ---- |
| YFe₁₁Ti, K (MJ/m³) | [−0.22](https://ouro.foundation/actions/01a1dae8-aa90-7fb4-a332-d863d8be140e) | [+0.19](https://ouro.foundation/actions/01a1dfd5-70a3-7331-881d-3663ce2f9d92) | [+0.84](https://ouro.foundation/actions/01a1dfd5-715c-7994-8251-e6b1adbdda82) | [+1.48](https://ouro.foundation/actions/01a1dfd5-712c-776c-a972-aa19e5a8c703) | [+1.78](https://ouro.foundation/actions/01a1dfd5-716d-7cc1-b8e9-f63304d4e4d3) | **[+2.02](https://ouro.foundation/actions/01a1dfd5-7169-794a-862b-b4ad3204410b)** |

At the measured moment (19.97 μB total, about 19.4 of spin) the route gives 2.02 against 2.07 measured. The measured cobalt series falls into place the same way. Single crystals of YFe₁₁₋ₓCoₓTi have their largest K₁ at one Co and a sharp fall beyond it (Nikitin et al., 2001); at its free moment our route instead climbs from −0.22 to [+1.01](https://ouro.foundation/actions/01a1dae8-aa58-7abb-a8c3-2613060ba043) to [+1.44](https://ouro.foundation/actions/01a1dae8-a978-731b-8bd8-52011fa8de87) for 0, 1 and 3 Co. With the moment held lower, the one-Co cell peaks at +2.08 and the three-Co cell falls (these two are bench values, not rerun through the route).

The same scan on the Zr cells, each on its DFT-relaxed geometry, moment held at a fraction of the cell's own:

| Fraction of the free moment | 1.00 (free) | 0.97 | 0.94 | 0.91 | 0.88 | 0.85 | 0.82 |
| --------------------------- | ----------- | ---- | ---- | ---- | ---- | ---- | ---- |
| ZrFe₁₁Ti | [+0.31](https://ouro.foundation/actions/01a1db26-ff65-7a69-88fb-f631bfd6a3b9) | [+0.56](https://ouro.foundation/actions/01a1dfd5-7145-7014-88ef-d571f7913ed8) | [+0.90](https://ouro.foundation/actions/01a1dfd5-70ca-7c30-b6b3-b80c526f0ff1) | [+1.37](https://ouro.foundation/actions/01a1dfd5-70dd-7c77-872b-28a140a076c7) | [+1.85](https://ouro.foundation/actions/01a1dfd5-7166-7107-916a-2fbccc226ec9) | [+2.27](https://ouro.foundation/actions/01a1dfd5-71b9-7af8-881d-b8672c3f8fa9) | [+2.39](https://ouro.foundation/actions/01a1dfd5-7157-7a41-a860-5333411f81df) |
| **ZrFe₁₀CoTi** | [+1.12](https://ouro.foundation/actions/01a1db3b-2682-792e-8fe2-98a3cdedc720) | [+1.24](https://ouro.foundation/actions/01a1dfd5-7165-77c0-be18-1ae59afb2c4e) | [+1.29](https://ouro.foundation/actions/01a1dfd5-7168-7e76-830c-a490d52f92a0) | [+1.32](https://ouro.foundation/actions/01a1dfd5-7140-732d-b52d-c6f5e16794f6) | [+1.35](https://ouro.foundation/actions/01a1dfd5-716c-7d8c-bb3e-8d1e792efce4) | [+1.33](https://ouro.foundation/actions/01a1dfd5-7169-794a-862b-b4ac4ca6f3bd) | [+1.24](https://ouro.foundation/actions/01a1dfd5-7164-7db8-90a4-2c55b6971c56) |
| ZrFe₉Co₂Ti | [+1.78](https://ouro.foundation/actions/01a1db26-ffbb-7496-9e7f-6617f899cc3d) | [+1.90](https://ouro.foundation/actions/01a1dfd5-716d-7cc1-b8e9-f632c95f12d5) | [+1.66](https://ouro.foundation/actions/01a1dfd5-717e-7a65-bfcd-9cf670b418f7) | [+1.13](https://ouro.foundation/actions/01a1dfd5-714c-7149-861f-311fcb9f68c4) | [+0.55](https://ouro.foundation/actions/01a1dfd5-7194-73d4-b400-756d8891d4a2) | [+0.06](https://ouro.foundation/actions/01a1dfd5-717d-79b1-9833-f433424657f6) | [−0.30](https://ouro.foundation/actions/01a1dfd5-715d-705d-822e-76b897bade7b) |
| ZrFe₈Co₄, Co on 8f | [+4.01](https://ouro.foundation/actions/01a1db15-8f67-7890-9375-410d80b92468) | [+3.79](https://ouro.foundation/actions/01a1dfd5-715b-75e1-8c81-4ddbe28fd6bf) | [+2.94](https://ouro.foundation/actions/01a1dfd5-7146-7596-9718-fc663529c8b6) | [+1.91](https://ouro.foundation/actions/01a1dfd5-7149-748d-8408-dd1f4be64b96) | [+0.99](https://ouro.foundation/actions/01a1dfd5-7142-7519-af39-c17eff40fb28) | [+0.48](https://ouro.foundation/actions/01a1dfd5-7156-7575-9d84-831572242978) | [+0.39](https://ouro.foundation/actions/01a1dfd5-716a-7860-9c27-97c5733eac50) |

Every cell traces a dome 2–4 MJ/m³ high, and composition decides where the cell's own moment sits on it. The iron-only cell sits on the high-moment side and climbs as the moment is lowered. The cobalt-rich cells sit at the top and fall: ZrFe₉Co₂Ti holds 1.1–1.9 MJ/m³ down to 0.91 of its moment and is gone by 0.85. ZrFe₁₀CoTi is flat: between 1.1 and 1.35 MJ/m³ over the whole range. PBE over-polarises YFe₁₁Ti by 10 % (all-electron) to 15 % (our pseudopotential) against experiment, and we do not know the true moment of a compound nobody has made, so the flat cell is the one whose reading we can stand behind. If cobalt brings the calculated moment closer to the true one, as the measured magnetisation of the Y series suggests, ZrFe₉Co₂Ti at about 1.7–1.9 MJ/m³ and 1.5 T is the stronger magnet; that rests on an assumption the flat cell does not need.

## What the hypotheses did

- **An unmeasured hard magnet is still sitting in the databases: refuted.** We ran the anisotropy route on the 23 iron-rich, rare-earth-free, uniaxial structure types within 25 meV/atom of the hull that the campaign had not measured. None reads 1 MJ/m³ or more. The largest easy-axis readings are tetragonal α-MoB-type FeB, [+0.82](https://ouro.foundation/actions/01a1dad4-1f8f-79ad-8c60-f7ebe77ee04f) at 1.04 T, and Fe₃B in the Fe₃P type, [+0.71](https://ouro.foundation/actions/01a1dad4-2228-7408-8f8e-11eff751cdd0), which experiment has as easy-plane.
- **Some element holds only the 8f site of ZrFe₁₂: refuted in practice.** Of 18 elements, only P and B prefer 8f by more than 150 meV, and both push the cell 70–80 meV/atom above the hull. Ti and Mo take 8i by about 0.55 eV and lower the hull from 30 to 21 meV/atom.
- **Our route reproduces the measured cobalt trend of Y(Fe,Co)₁₁Ti: refuted at the free moment, as predicted**, and explained by the moment.
- **YFe₁₁Ti reads easy-plane because its moment is too high: supported.** Predicted +1.3 to +2.8 at the measured moment; found +2.02 on the route (+1.65 to +2.04 on the bench, depending on smearing).
- **Cobalt holds the 8f site in ZrFe₈Co₄: refuted, as predicted.** On PBE the [all-8j arrangement](https://ouro.foundation/actions/01a1db55-01c5-793a-aeab-303b85b973e0) is 25 meV per Co above all-8f, far under the thermal energy of an anneal, and with Co on 8j the same composition reads [−0.33](https://ouro.foundation/actions/01a1daec-4e21-7c45-a2a5-8144012adfed) (on an Orb-relaxed cell). The 4.0 MJ/m³ cell is one ordering of an alloy that would not hold it.

## What changed

- Anisotropy in the 1:12 family is set by how full the minority band is. A reading at the calculation's own moment can be wrong in sign for iron-only cells and wrong in ranking across a cobalt series.
- Cobalt is not a lever for anisotropy here. At the right moment the first cobalt per formula unit is already the top, as the measured Y crystals show; what cobalt buys is Curie temperature and insensitivity to the moment.
- Mo on 8i counts as two Co in electron count: ZrFe₁₁Mo reads [+1.98](https://ouro.foundation/actions/01a1dada-a33b-7ecd-892c-4dd7d6018fb6) at its free moment (Orb-relaxed cell) and falls with the moment like the two-Co cell.
- If the true moment of ZrFe₁₁Ti is where YFe₁₁Ti's is, the cobalt-free cell is the better magnet (1.8–2.4 MJ/m³ near 1.4 T, at 5.48 $/kg). A published calculation gives it 1.1–2.5 MJ/m³ depending on functional and Ti arrangement (J. Magn. Soc. Jpn. 48, 2407R001, 2024), which is the same spread seen from the other side.

## What this is not

Nothing here has been made. The only rare-earth-free 1:12 phase reported is Zr(Fe,Si)₁₂, and it is metastable; ZrFe₁₀CoTi sits 26 meV/atom above the hull where that compound sits at 4 on the same scale. The candidate's reading is one Ti and Co arrangement in a small basis. With Ti along one in-plane axis the two in-plane directions differ, and c is easier than the nearer one by only 0.69 MJ/m³. "New composition" means absent from Materials Project, Alexandria, GNoME and COD and not found in three literature searches.

## Instruments

- The DFT routes refuse scratch uploads, so the first batch of 23 runs failed at once and was rerun from private files.
- A DFT relaxation makes the moment problem worse: the over-polarised cell expands by 3.4 % and its moment grows. The Orb-relaxed ZrFe₁₁Ti reads +0.84; the DFT-relaxed one +0.31.
- The route had no fixed-moment option when this was first posted. It has one now, and the two tables above were rerun through it on 2026-11-15. The first version of the Zr table listed the ZrFe₈Co₄ row against the wrong fractions: its bench scan stepped by 2 %, not 3 %, so its values fell faster than the row showed. The row is now measured at the fractions in the header.

## Next

1. Rerun ZrFe₁₀CoTi and ZrFe₁₁Ti through the fixed-moment route in the larger basis, on a second Ti and Co arrangement.
2. Makeability first: mixed Si + Ti stabilisers with one Co, placed against the made ZrFe₁₀Si₂, and a moment scan of ZrFe₁₀Si₂ itself as a second calibration.
3. Re-read the campaign's other instrument failures (Fe₃P, Fe₅SiB₂, the Laves calibration) with the moment in hand.
4. Scan the database pool between 25 and 75 meV/atom, which the campaign's own stability ceiling admits and nobody has screened.

## Assets in this post

- [ZrTiFe10Co (Imm2) - DFT relaxed](https://ouro.foundation/files/mmoderwell/zrtife10co-imm2-dft-relaxed) (file, id `e0d3ec14-05c6-425c-a71e-97bd19a0853e`)
- [DFT structure relaxation](https://ouro.foundation/routes/mmoderwell/dft-structure-relaxation) (route, id `d180f520-fc88-48b7-9eca-3d123f5cf680`)
- [Magnetic anisotropy energy](https://ouro.foundation/routes/mmoderwell/magnetic-anisotropy-energy) (route, id `331d9faf-4b44-4679-a958-5a9a816036e1`)
- [ZrTiFe11 (Imm2) - DFT relaxed](https://ouro.foundation/files/mmoderwell/zrtife11-imm2-dft-relaxed) (file, id `c659e63d-cf92-457f-afd2-244a33774e19`)
- [ZrTiFe9Co2 (Imm2) - DFT relaxed](https://ouro.foundation/files/mmoderwell/zrtife9co2-imm2-dft-relaxed) (file, id `0b445ab3-2481-4dab-bd26-1adf05f39f24`)
- [Zr(Fe2Co)4 (I4/mmm) - DFT relaxed](https://ouro.foundation/files/mmoderwell/zrfe2co4-i4-mmm-dft-relaxed) (file, id `f85e92a7-1875-47a8-a1d4-d6e692929633`)
