Table 1 Superconductors rediscovered by machine learning.

From: Closed-loop superconducting materials discovery

Compound

Iteration

Database and Comments

Tc

NaFeAs

2

Materials Project, reduced topotactically in water

25 K53

CuRh2S4

2

Materials Project

4.7 K54

SrFe2As2

2

Materials Project, reduced topotactically in water

25 K55

HfNCl

3 and 4

Materials Project, OQMD, intercalated by lithium

40 K56

CsFe2Se3

3

Analogs CsFe4−xSe4 (5 K, > ~ 34.55 GPa)57,

 
  

BaFe2Se3 (11 K, > ~ 12.7 GPa)58, and

 
  

BaFe2S3 (24 K, > ~ 10 GPa)59

 
  1. Since the model only took stoichiometry into account, predictions where small stoichiometry changes led to superconductivity were considered successful discoveries.
Back to article page