GGen ran 1,000 trials across 100 Co-Fe-Ta stoichiometries in 19.8 min, keeping 970 relaxed structures. 68 of them landed within 150 meV/atom of the convex hull, 66 of which the database had not seen before. The leading generated candidate is CoFe6 in R-3m (#166), which sits 1 meV/atom above the hull (known).
GGen ran 1,000 trials across 100 Co-Fe-Ta stoichiometries in 19.8 min, keeping 970 relaxed structures. 68 of them landed within 150 meV/atom of the convex hull, 66 of which the database had not seen before. The leading generated candidate is CoFe6 in R-3m (#166), which sits 1 meV/atom above the hull (known).
Metric | Value |
|---|---|
Chemical system | Co-Fe-Ta |
Stoichiometries | 100 |
Trials per stoichiometry | 10 |
Total trials | 1,000 |
Structures kept | 970 across 102 formulas |
Failed stoichiometries | 0 |
Phases on the hull | 0 |
Phases within 150 meV/atom | 68 |
New compositions | 64 |
New polymorphs | 2 |
Best hull distance | 1 meV/atom (CoFe6) |
Wall time | 19.8 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 | 12,274 |
Generated by GGen | 12,036 |
From Alexandria | 183 |
From the Materials Project | 55 |
Hull distances are computed against every Co-Fe-Ta 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 289 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.
CoFe6 in R-3m (#166), 1 meV/atom above the hull, -8.281 eV/atom, 21 atoms.
GGen Co-Fe-Ta candidate: R-3m (#166), 1 meV/atom above the hull, -8.281 eV/atom, 21 atoms
Co2Fe8 in C2/m (#12), 1 meV/atom above the hull, -8.216 eV/atom, 20 atoms.
GGen Co-Fe-Ta candidate: C2/m (#12), 1 meV/atom above the hull, -8.216 eV/atom, 20 atoms
Fe15Ta in C2/m (#12), 9 meV/atom above the hull, -8.679 eV/atom, 32 atoms.
GGen Co-Fe-Ta candidate: C2/m (#12), 9 meV/atom above the hull, -8.679 eV/atom, 32 atoms
The hull is drawn from every Co-Fe-Ta 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-Ta structure in the GGen database
Stage | Seconds | Share |
|---|---|---|
Candidate generation | 1154.0 | 97% |
Candidate relaxation gpu | 987.3 | 83% |
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 |
382 |
Subsystems covered | 7 |
Phases on the hull | 16 |
Phases within 150 meV/atom | 289 |
Exploration runs | 1 |
Last explored | 2026-08-25T06:43:02 |
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 | no | -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 | Ta2Co | I4/mcm (#140) | tetragonal | 0 | yes | reference | MP | no | -10.527 | 6 | 15.6 |
12 | Ta6Fe7 | R-3m (#166) | trigonal | 0 | yes | reference | MP | no | -10.227 | 13 | 14.0 |
13 | Ta7Co6 | R-3m (#166) | trigonal | 0 | yes | reference | MP | no | -9.922 | 13 | 14.4 |
14 | Ta7Fe6 | R-3m (#166) | trigonal | 0 | yes | reference | MP | no | -10.461 | 13 | 14.7 |
15 | TaCo3 | R-3m (#166) | trigonal | 0 | yes | reference | MP | no | -8.527 | 12 | 12.0 |
16 | TaFe2 | P6_3/mmc (#194) | hexagonal | 0 | yes | reference | MP | no | -9.782 | 12 | 12.8 |
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 | CoFe24 | C2/m (#12) | monoclinic | 0 | no | new composition | GGen | no | -8.393 | 50 | 11.5 |
20 | Co4Fe8 | P2/m (#10) | monoclinic | 0 | no | new polymorph | GGen | no | -8.056 | 12 | 11.6 |
21 | CoFe20 | C2/m (#12) | monoclinic | 0 | no | new composition | GGen | no | -8.385 | 42 | 11.5 |
22 | Fe9Co7 | Pm-3m (#221) | cubic | 0 | no | reference | MP | no | -7.920 | 16 | 11.5 |
23 | CoFe11 | P2/m (#10) | monoclinic | 0 | no | new composition | GGen | no | -8.347 | 12 | 11.5 |
24 | CoFe13 | C2/m (#12) | monoclinic | 0 | no | new polymorph | GGen | no | -8.360 | 28 | 11.5 |
25 | Co2Fe2 | Pm-3m (#221) | cubic | 0 | no | known | GGen | no | -7.833 | 2 | 11.5 |
176.0 |
15% |
Terminal generation | 32.6 | 3% |
Hull and stability | 1.4 | 0% |
Setup and load | 1.1 | 0% |
Enumeration | 0.3 | 0% |
Finalization | 0.0 | 0% |
Total | 1189.4 | 100% |
Crystal systems |
tetragonal, hexagonal, orthorhombic |
Min element fraction | {'Fe': 0.5} |
Max element fraction | {'Ta': 0.2} |
Require all elements | False |
Skip existing formulas | False |
Near-hull cutoff (eV/atom) | 0.15 |
Relaxation max steps | 400 |
Relaxation optimizer | fire |
Same treatment I gave your Co-Fe-Hf run: the three best structures from this exploration through the magnet 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 |
|---|---|---|---|
CoFe6 (R-3m, 1 meV/atom) | 1 meV/atom | 2.39 T (1.90 MA/m) | 980 K |
Co2Fe8 (C2/m, 1 meV/atom) | 1 meV/atom | 2.36 T (1.88 MA/m) | 967 K |
Fe15Ta (C2/m) | 9 meV/atom | 2.17 T (1.72 MA/m) | 816 K |
What catches my eye is Fe15Ta rather than the two Fe-Co binaries. The binaries are essentially bcc-Fe-like — moments 2.4-2.7 µB on Fe, ~1.4 µB on Co, Ms right where Slater-Pauling counting puts them, and Curie screens near pure iron's 1043 K. Real magnets, but nothing you couldn't guess from the alloy rulebook. Fe15Ta is the one doing something structural: Ta carries ~0.02 µB (same as Hf yesterday — the refractory element is there for stability, not magnetization), yet the Fe sublattice holds 1.9-2.5 µB/site and the phase sits only 9 meV/atom above a hull that includes the known TaFe2 Laves phase. That's a rare-earth-free pattern worth naming: a non-magnetic refractory stabilizer that barely dilutes the magnetization. The 816 K Curie screen with 2.17 T Ms is comfortably above room temperature, though as always these are FM-assumed and heuristic — CHGNet cannot tell us FM from AFM, and the Curie screen is a regression, not an exchange calculation.
Receipts: moments and Ms for CoFe6, Co2Fe8, and Fe15Ta; Curie screens here, here, and here.
If you want, I'll queue DFT SCF → MAE → TB2J on Fe15Ta — between it and yesterday's Fe15Hf and CoFe14Hf, that would give us a small matched set of "refractory-stabilized Fe-rich" candidates to see whether any of them develop anisotropy, which is the property none of these screens can see.