Fig. 5: Experimental and ML guided explanation of degradation phenomenon. | Nature Communications

Fig. 5: Experimental and ML guided explanation of degradation phenomenon.

From: Co-expression of multi-genes for polynary perovskite electrocatalysts for reversible solid oxide cells

Fig. 5

a Relationship between degradation rate and Ln-O/Ba-O bond length. b Surface A/B and Ba/Pr ratios of PBC before and after stability test. c Thermal stability study of PBC after 1 × 106 Monte Carlo (MC) steps at 973 K. d DFT calculated formation energy of BaO. e The surface A/B and Ba/Pr ratios of PLNSY before and after stability test. f Thermal stability study of PLNSY after 1 × 106 MC steps at 973 K. g Thermal cycling stability of PLNSY. h Long-term stability test of PLNSY in a BZCYYb-supported symmetric cell configuration at 600 oC and 3% wet air. i Correlation matrix for atomic theoretical features and bond length information. j Performance of the linear regression model, with the horizontal axis representing the number of input features. The performance indicates the best model optimized using modified recursive feature elimination. k Normalized coefficients, derived from the linear regression coefficients without bond length information, expressed as percentages of the sum of absolute coefficients.

Back to article page