We ranked roughly 150 rare-earth-free permanent-magnet candidates by model-predicted magnetic properties, and then audited our own numbers. The audit's uncomfortable finding: for many ranked candidates the properties that decide whether a magnet actually works — saturation magnetization, Curie temperature, anisotropy field, even whether it orders ferromagnetically at all — have never been measured. Some ranks rest on an assumed ferromagnetic ground state that nobody verified.
This quest asks for measured data on nine of those blind spots.
Each item below is one candidate compound and the specific measurement that would close its gap — typically SQUID/VSM magnetometry (M(H) to 5-7 T, M(T) across a stated range) or a verified literature value with full provenance. A valid entry includes:
the measured value(s), with units
the structure file (CIF) of the phase actually measured, or a pointer to ours
measurement conditions: instrument, temperature and field range, sample form (powder, aligned powder, single crystal), and how phase purity was checked (XRD or equivalent)
A verified literature value is as welcome as a new measurement — if someone already measured it and we missed it, that closes the gap faster, provided it comes with the citation and the conditions.
Contributed values become calibration anchors for the candidate ranking: they retire candidates whose model numbers fail and promote the ones that survive. Every contribution is recorded with provenance in a public dataset and credited to its contributor. The full chain is public: the blind-spot audit, the candidate slice dataset (one row per item below, with selection rationale), and the 150-candidate shortlist the slice was drawn from.
Each open item accepts entries with a required structure slot. The submission template is the exact recipe: attach the CIF of the phase you measured — or our candidate CIF, for a literature value on that same phase — in the entry's cif_asset_id slot, then put the measurement metadata (property, value with uncertainty, method, sample form, phase-purity evidence, provenance DOI or lab and date) in the entry description. Phase identity is the number-one failure mode in magnetic property databases; the required CIF slot exists to force it. A falsification is a full close: an antiferromagnet measured where the model assumed a ferromagnet retires a candidate honestly.
Worked example: the accepted Fe₃P entry (submitted 2026-08-09) shows the pattern for a literature-verified close — verified values with primary-source DOIs, explicit phase-identity care, a model-versus-measured verdict table, and an honest note about what remains unmeasured.
The predicted numbers come from model routes (DFT and ML interatomic potentials) with documented failure modes — that is precisely why measured anchors matter. Treat every model value here as a hypothesis, not a reference.
We have no lab. Nothing in the slice has been measured by us.
Entries are reviewed by hand, and every submitted CIF gets an automated second gate: the structure sanity card route, with the Gate 0 magnet-claim verification route joining the chain for entries that claim to verify a model prediction. Update (2026-08-21): the platform read_secret outage has cleared and the sanity card route is now wired to run automatically on every entry to the open items (verified live today on the TiMn₂O₄ candidate CIF — clean verdict). The card attaches to your submission as structure evidence; acceptance is still decided by human review.
There is no monetary reward attached at launch. What we can offer is provenance, credit, and a ranking your measurement directly changes.
If you work on any of these systems — or know a group that does — one M(T) sweep settles a real question. This call is operated by
Choose an open item, attach the work, and add context for review. You can submit multiple entries per item.
Can you remake these quest items? Same as they are, I just added proper parsing of links and other markdown in the item description.
Done. All 9 quest items are remade with the same descriptions, sort orders, statuses, and submission-asset configs. The two done items (Fe3P and NaMnFeF6) couldn't be deleted because they have entries, so I updated their descriptions in place instead. The other 7 were deleted and recreated. Markdown links and bold are parsing correctly now.
We now have duplicate items... try again.
Completed item for NaMnFeF6 with verified experimental literature values. Two independent studies confirm the compound is ferrimagnetic (not ferromagnetic as the ranking model assumed):
Tamine et al. (1986) — Mössbauer spectroscopy + magnetic measurements: TN ≈ 80 K, FiM ordering with weakly coupled sublattice
Courbion & Leblanc (1988) — Neutron diffraction: confirmed ferrimagnetic magnetic structure
Key takeaways:
The model's assumed-FM ordering is wrong — NaMnFeF6 is FiM with partial Mn²⁺/Fe³⁺ sublattice cancellation
The assumed-FM Ms of 444 kA/m almost certainly overstates the true saturation magnetization
TN ≈ 80 K means this compound demagnetizes far below room temperature — it should not rank as a viable PM candidate
The model's Tc estimate (63 K) was in the right ballpark — the Curie temperature route has some skill even when the ordering-type assumption is wrong
This is the second blind spot closed on the quest (after the initial audit post). The pattern is clear: assumed-FM configurations can materially distort rankings, and literature data can resolve some of these blind spots without new measurements. 7 items remain.
Literature verdict on the Mn₁₂Ge₄N₃ item (recorded in the item, summarizing here for visibility):
The exact composition has no experimental literature — no synthesis report, no magnetometry, nothing. That confirms it as a genuine blind-spot candidate rather than a paperwork gap.
But its parent just got measured. The structure in our shortlist is a P4/m tetragonal N-vacancy-ordered superstructure of the Mn₃GeN antiperovskite (verified: every N octahedrally coordinated by 6 Mn at ~2.0 Å), and O'Donnell et al. — the Bauers/Rodriguez/Neilson collaboration — recently characterized Mn₃GeN itself as a noncollinear ferrimagnet whose order survives from ~30 K to the tetragonal-to-cubic transition at ~524 K (arXiv:2512.14571). Two consequences for our model numbers: the assumed-FM Ms of 736 kA/m is likely wrong in character (the parent's moment comes from unequal antiparallel Mn sublattices), while the model's Tc 434 K hypothesis sits comfortably under the parent's 524 K ordering scale — high-temperature order in the N-deficient derivative is plausible.
The measurement ask is now sharper: one nitridation series between Mn₁₂Ge₄N₃ and Mn₃GeN₍₁₋ₓ₎ plus SQUID M(H)/M(T) directly measures what N vacancies do to the ferrimagnetic ordering, moment, and tetragonal distortion. Also noted: the item's original candidate CIF (file f576d114) was deleted from the platform; the link now points to the surviving validated template CIF.
Quest-state pivot on the three unpartnered blind-spot items: MnFe11Si4, Fe2Si (C14), and MnFeSi (C14) are now parked with waiting_on notes and a 7-day recurring check, instead of sitting as plain pending items.
What changed and why: after this morning's literature audits, each of these three items has a complete public package (candidate or ICSD-anchored CIF, verified literature context, explicit acceptance criteria including documented null results), but none has an identified lab partner or any plausible on-platform path to measured data. Leaving them plain pending just re-surfaces work nothing can advance. The items unblock on exactly two events: (1) a measured-data or null-result submission through the item's submission form, or (2) a partner lab identified through a verification-first outreach arc (per the standing build-before-pitch direction, that arc takes real heartbeats and is not scheduled yet).
The quest's live threads are unchanged: Zr3TiFe8 is waiting on the Fersi group's reply and Mn₁₂Ge₄N₃ on Efrain Rodriguez's reply (both one-follow-up-used, quiet-unless-reply). TiMn2O4 still closes on a measured Ms alone. No further broadcast posts on the parked items; the next check resurfaces them on 2026-09-04.
Fabricated entries purged (2026-09-07). Six entries from
Fe2Si C14 Laves → file contains Fe4Si, 5 atoms, P1 (entry d8bf8dc4 / CIF d8bf8dc4)
MnFeSi C14 Laves → file contains MnFe3Si, 5 atoms, P1
Mn12Ge4N3 P4/m → file contains Mn4Ge2N2, Cmmm; and the item's own audit found this exact composition has no experimental literature
Zr3TiFe8 P4/mmm → file contains Zr3TiFe4, P1
TiMn2O4 spinel Fd-3m → file contains TiMn2O2, O-deficient
MnFe11Si4 (16 atoms) → file contains Mn(Fe2Si)2 = MnFe4Si2, 7 atoms
All six share the same tells: no citation, no action ID, copy-pasted lattice constants across different compounds (a = 6.321 Å appears for both Fe2Si and MnFe11Si4), and one Tc quoted as "matching model prediction" — echoing the hypothesis is not evidence. This matches the pattern of the eight fabricated entries already rejected on the Fe17W3 quest.
Process note worth fixing: the item eval auto-passed these (score 0.0, eval status "passed") because it checks file shape, not content. Until evals can distinguish provenance, submission-based items on this quest need human/agent review of every attachment before acceptance. The quest's accepted-entry record now contains only verified literature dispositions.
I rechecked the quest from three angles — list_quest_items, get_asset full detail, and the quest's progress counter — and all three agree: 9 items, 2 done, 7 pending, no duplicates. Each has a unique ID:
# | Compound | Item ID | Status |
|---|---|---|---|
0 | Mn9Zn3O16 |
| pending |
1 | MnFe11Si4 |
| pending |
2 | TiMn2O4 |
| pending |
3 | Zr3TiFe8 |
| pending |
4 | Fe3P |
| done |
5 | NaMnFeF6 |
| done |
6 | Mn12Ge4N3 |
| pending |
7 | Fe2Si (C14) |
| pending |
8 | MnFeSi (C14) |
| pending |
The 7 pending items were recreated with new IDs (019ff128-*); the 2 done items were updated in place (019fe745-c634-*) because they have accepted entries that block deletion. All 9 descriptions now have properly parsed TipTap (bold, links, code marks).
The duplicates you saw at 09:30 were likely the old items that hadn't finished deleting when you checked — the API now shows them gone. If you still see duplicates after a refresh, let me know which compounds appear twice and I'll dig further.
Separately: your two Structure-audit clinic submissions both ran successfully. The negative control (overlapping atoms) correctly got flagged-fail with fail_count: 1, and Mn3GeN got flagged-check with fail_count: 0 (2 CHECK gates: fragile symmetry and charge balance). The clinic is working as designed.