Rational Design for Efficient Bifunctional Oxygen Electrocatalysts by Artificial Intelligence

High-performance
bifunctional electrocatalysts that simultaneously
and efficiently catalyze oxygen reduction reaction (ORR) and oxygen
evolution reaction (OER) have become the bottleneck and main challenge
for rechargeable zinc–air batteries. The components need to
be comprehensively designed to enhance the formation possibility for
the respective activities of ORR and OER. Nevertheless, the elements
and types of chemical bonds generate an almost infinite space of potential
candidates. In this work, we proposed a rational design strategy for
efficient bifunctional oxygen electrocatalysts by a data-driven method.
According to the inferred ΔE from E10 and E1/2 machine learning
models among all the bond combinations, a bond combination of C–N,
C–C, Fe–N, Ru–O, and C–P was considered
with superior possibility to have bifunctional activity owing to its
highest occurrence frequency. The experimental results further confirmed
that this component and bonding method indeed have ORR/OER bifunctional
activity, which has not been reported yet. This strategy brings novel
and efficient insights for bifunctional electrocatalyst design from
a huge potential exploration space.