A Multiobjective Closed-loop Approach Towards Autonomous Discovery of Electrocatalysts for Nitrogen Reduction

29 May 2023, Version 2
This content is a preprint and has not undergone peer review at the time of posting.


Electrocatalyst discovery is an inherently multiobjective challenge that can benefit from closed-loop approaches towards acceleration. However, previous computational closed-loop efforts for electrocatalysis often focus on a single objective to be optimized. Here, we demonstrate a multiobjective closed-loop strategy towards identifying single-atom alloy (SAA) electrocatalysts for nitrogen reduction considering activity, stability, and cost. Candidates were autonomously selected via a multiobjective scoring approach, as implemented in our AutoCat software, and evaluated using a high-throughput density functional theory pipeline. We discuss the implications of our formulation of the candidate scoring system and show its ability to efficiently explore the SAA design space. We also propose a multiobjective ranking scheme balancing the three target metrics to prioritize further investigation of the top-performers identified here, Zr$_1$Cr, Hf$_1$Cr, Ag$_1$Re, Au$_1$Re, and Ti$_1$Fe. The inclusion of Hf in the top-ranked candidates is of particular interest as it is more commonly found within the context of molecular catalysts.


sequential learning
autonomous discovery
multiobjective optimization

Supplementary materials

Supplementary Information
Tabulated partial scores for candidates that fall within the activity window, descriptions related to data and scripts for reproducibility

Supplementary weblinks


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.