Campaign statistics for the chemical systems studied in this work. Gross discovery represents the total number of new materials within 0.2 eV per atom of the convex-hull at the end of the campaign, prior to post-campaign structure analysis. Duplicate discovery column lists how many of the structures among the gross discoveries were found to be duplicates of others. We further checked how many of the discovered structures relaxed into ICSD-derived entries in OQMD, listed as experimental in the table. Final column lists the total number (383) of explicitly unique, non-experimental structures.
System | Agent requests | DFT received | Gross discovery | Duplicate discovery | Experimental | Unique discovery |
---|---|---|---|---|---|---|
B–Fe | 152 | 146 | 32 | 28 | 3 | 19 |
C–Fe | 91 | 81 | 37 | 14 | 2 | 26 |
Cl–Os | 101 | 70 | 13 | 1 | 1 | 11 |
Fe–N | 66 | 47 | 17 | 10 | 1 | 10 |
Fe–S | 158 | 133 | 22 | 23 | 6 | 14 |
Fe–V | 147 | 130 | 53 | 30 | 1 | 32 |
Mn–S | 99 | 73 | 29 | 5 | 2 | 22 |
Ni–S | 220 | 191 | 94 | 30 | 6 | 66 |
O–V | 66 | 52 | 23 | 5 | 3 | 17 |
S–V | 174 | 129 | 31 | 22 | 3 | 20 |
Al–Sc–Ti | 216 | 199 | 115 | 32 | 0 | 90 |
Bi–Ca–P | 220 | 157 | 37 | 10 | 0 | 27 |
Ca–Mo–P | 73 | 53 | 6 | 2 | 0 | 4 |
Cu–O–Rh | 123 | 72 | 17 | 4 | 1 | 14 |
Hf–K–P | 89 | 63 | 8 | 4 | 0 | 7 |
Hf–K–S | 38 | 23 | 4 | 0 | 0 | 4 |