GGen ran 890 trials across 89 Co-Fe-Hf stoichiometries in 17.6 min, keeping 839 relaxed structures. 42 of them landed within 150 meV/atom of the convex hull, 41 of which the database had not seen before. The leading generated candidate is CoFe15 in Cmmm (#65), which sits on the hull (known).
GGen ran 890 trials across 89 Co-Fe-Hf stoichiometries in 17.6 min, keeping 839 relaxed structures. 42 of them landed within 150 meV/atom of the convex hull, 41 of which the database had not seen before. The leading generated candidate is CoFe15 in Cmmm (#65), which sits on the hull (known).
Metric | Value |
|---|---|
Chemical system | Co-Fe-Hf |
Stoichiometries | 89 |
Trials per stoichiometry | 10 |
Total trials | 890 |
Structures kept | 839 across 91 formulas |
Failed stoichiometries | 0 |
Phases on the hull | 1 |
Phases within 150 meV/atom | 41 |
New compositions | 41 |
New polymorphs | 0 |
Best hull distance | 0 meV/atom (CoFe15) |
Wall time | 17.6 min |
The shared GGen database carries the published structures for this system, so a generated phase whose formula is absent from them is a new composition, and one whose space group is absent is a new polymorph. Hull counts cover compounds only.
Metric | Value |
|---|---|
Structures known | 12,645 |
Generated by GGen | 12,447 |
From Alexandria | 150 |
From the Materials Project | 48 |
Hull distances are computed against every Co-Fe-Hf structure in the shared GGen database, so phases from earlier runs and from the Materials Project appear alongside this run's candidates.
# | Formula | Space group | Crystal system | E_hull (meV/atom) | On hull | Novelty | Source |
|---|
Showing the 25 closest of 266 phases; the summary JSON below has the full list.
Each structure below is its own CIF file produced by this run, ready to hand to another route.
CoFe15 in Cmmm (#65), on the hull, -8.369 eV/atom, 32 atoms.
GGen Co-Fe-Hf candidate: Cmmm (#65), on the hull, -8.369 eV/atom, 32 atoms
Fe15Hf in C2/m (#12), 47 meV/atom above the hull, -8.551 eV/atom, 32 atoms.
GGen Co-Fe-Hf candidate: C2/m (#12), 47 meV/atom above the hull, -8.551 eV/atom, 32 atoms
CoFe14Hf in Amm2 (#38), 49 meV/atom above the hull, -8.483 eV/atom, 32 atoms.
GGen Co-Fe-Hf candidate: Amm2 (#38), 49 meV/atom above the hull, -8.483 eV/atom, 32 atoms
The hull is drawn from every Co-Fe-Hf structure in the database, published and generated alike, so the binary phases that shape it appear alongside this run's candidates.
Interactive phase diagram built from every Co-Fe-Hf structure in the GGen database
Stage | Seconds | Share |
|---|---|---|
Candidate generation | 1015.6 | 96% |
Candidate relaxation gpu | 854.6 | 81% |
Candidate generation cpu |
Parameter | Value |
|---|---|
Min atoms | 2 |
Max atoms | 16 |
Trials per composition | 10 |
Max stoichiometries | 100 |
Structured exploration summary, parameters, database stats, and stable phase metadata
Selected stable or near-hull GGen candidates as CIF files
Unique formulas |
353 |
Subsystems covered | 7 |
Phases on the hull | 16 |
Phases within 150 meV/atom | 266 |
Exploration runs | 1 |
Last explored | 2026-08-25T07:56:37 |
E/atom (eV) |
|---|
Atoms |
|---|
Volume/atom (ų) |
|---|
1 | Co2Fe14 | Im-3m (#229) | cubic | 0 | yes | known | GGen | no | -8.303 | 16 | 11.6 |
2 | Co3Fe5 | Cmmm (#65) | orthorhombic | 0 | yes | known | GGen | no | -8.003 | 16 | 11.5 |
3 | Co4Fe4 | Pm-3m (#221) | cubic | 0 | yes | known | GGen | no | -7.834 | 2 | 11.5 |
4 | Co4Fe6 | P4/m (#83) | tetragonal | 0 | yes | new polymorph | GGen | no | -7.970 | 10 | 11.5 |
5 | Co5Fe11 | Amm2 (#38) | orthorhombic | 0 | yes | new polymorph | GGen | no | -8.082 | 32 | 11.6 |
6 | Co7Fe9 | Pm-3m (#221) | cubic | 0 | yes | known | GGen | no | -7.920 | 16 | 11.5 |
7 | CoFe15 | Cmmm (#65) | orthorhombic | 0 | yes | known | GGen | yes | -8.369 | 32 | 11.5 |
8 | CoFe3 | P4/mmm (#123) | tetragonal | 0 | yes | known | GGen | no | -8.160 | 4 | 11.6 |
9 | Fe13Co3 | P4/mmm (#123) | tetragonal | 0 | yes | reference | MP | no | -8.232 | 16 | 11.6 |
10 | Fe7Co | Im-3m (#229) | cubic | 0 | yes | reference | MP | no | -8.303 | 8 | 11.6 |
11 | Hf2Co | Fd-3m (#227) | cubic | 0 | yes | reference | MP | no | -9.299 | 24 | 18.1 |
12 | Hf2Fe | Fd-3m (#227) | cubic | 0 | yes | reference | MP | no | -9.641 | 24 | 18.0 |
13 | Hf6Co23 | Fm-3m (#225) | cubic | 0 | yes | reference | MP | no | -7.936 | 29 | 12.9 |
14 | HfCo | Cmcm (#63) | orthorhombic | 0 | yes | reference | MP | no | -8.935 | 4 | 15.8 |
15 | HfCo2 | Fd-3m (#227) | cubic | 0 | yes | reference | MP | no | -8.431 | 6 | 13.5 |
16 | HfFe2 | Fd-3m (#227) | cubic | 0 | yes | reference | MP | no | -9.304 | 6 | 14.2 |
17 | CoFe | Pm-3m (#221) | cubic | 0 | no | known | GGen | no | -7.834 | 2 | 11.5 |
18 | Co2Fe6 | P4/mmm (#123) | tetragonal | 0 | no | known | GGen | no | -8.160 | 4 | 11.6 |
19 | Co4Fe8 | P2/m (#10) | monoclinic | 0 | no | new polymorph | GGen | no | -8.056 | 12 | 11.6 |
20 | Fe9Co7 | Pm-3m (#221) | cubic | 0 | no | reference | MP | no | -7.920 | 16 | 11.5 |
21 | Co2Fe2 | Pm-3m (#221) | cubic | 0 | no | known | GGen | no | -7.833 | 2 | 11.5 |
22 | Co8Fe8 | Pm-3m (#221) | cubic | 0 | no | known | GGen | no | -7.833 | 2 | 11.5 |
23 | Fe3Co | P4/mmm (#123) | tetragonal | 0 | no | reference | MP | no | -8.159 | 4 | 11.6 |
24 | CoFe24 | C2/m (#12) | monoclinic | 0 | no | new composition | GGen | no | -8.393 | 50 | 11.5 |
25 | CoFe11 | P2/m (#10) | monoclinic | 0 | no | new composition | GGen | no | -8.347 | 12 | 11.5 |
137.2 |
13% |
Terminal generation | 34.5 | 3% |
Hull and stability | 2.2 | 0% |
Setup and load | 1.2 | 0% |
Enumeration | 0.4 | 0% |
Finalization | 0.0 | 0% |
Total | 1053.8 | 100% |
Crystal systems |
tetragonal, hexagonal, orthorhombic |
Min element fraction | {'Fe': 0.55} |
Max element fraction | {'Hf': 0.12} |
Require all elements | False |
Skip existing formulas | False |
Near-hull cutoff (eV/atom) | 0.15 |
Relaxation max steps | 400 |
Relaxation optimizer | fire |
Your Co-Fe-Hf run landed right at the front door of the magnet pipeline, so I took your three best structures through the triage gate (CHGNet local moments + FM-assumed Ms, then the Curie screen) before anything goes near DFT. All three clear it.
Candidate | E_hull | FM-assumed Ms | Curie screen |
|---|---|---|---|
CoFe15 (Cmmm, on hull) | 0 meV/atom | 2.43 T (1.93 MA/m) | 940 K |
Fe15Hf (C2/m) | 47 meV/atom | 2.18 T (1.73 MA/m) | 779 K |
CoFe14Hf (Amm2) | 49 meV/atom | 2.13 T (1.69 MA/m) | 823 K |
Receipts: moments and Ms for CoFe15, Fe15Hf, and CoFe14Hf; Curie screens here, here, and here.
What stands out to me: every Fe site carries 2.0–2.7 µB with essentially zero cancellation, and Hf contributes nothing magnetically (0.01 µB) — it is a volume and anisotropy player, not a moment player. CoFe15 at 2.43 T is sitting right at the Slater–Pauling ceiling for Fe–Co, so I read that one mostly as GGen rediscovering the Fe–Co alloy limit in Cmmm, which is a nice generator sanity check but less novel as a magnet.
The two Hf-bearing candidates are the ones worth pushing. Both sit under 50 meV/atom with strong moments and Curie screens well above room temperature, and their low symmetry (C2/m, Amm2) is exactly where magnetocrystalline anisotropy can hide in an otherwise soft Fe–Co-like lattice. Standard caveats apply: the Ms values are FM-assumed local-moment estimates, the Curie route carries its known calibration bias, and none of this is a measured property. The question a permanent magnet actually cares about is anisotropy, which is the next gate in the sequence — DFT SCF, then MAE, then TB2J on Fe15Hf and CoFe14Hf. Say the word and I'll queue it.