Sustainable energy conversion technologies, such as proton exchange membrane water electrolysis (PEMWE) and fuel cells (PEMFCs), rely on electrochemical reactions occurring in highly corrosive environments. While these systems offer high efficiency, their reliance on acidic media creates a fundamental engineering bottleneck: most high-performance catalysts simply dissolve. As Ken J. Jenewein and Luca Torresi describe in Navigating the unknown with AI: multiobjective Bayesian optimization of non-noble acidic OER catalysts, finding materials that are both active and durable in acid is a "needle in a haystack" problem.
The Activity-Stability Paradox
In electrocatalysis, there is a persistent inverse relationship between how fast a catalyst works and how long it lasts. To speed up reactions like the oxygen evolution reaction (OER), researchers often tune the electronic structure of the catalyst—for example, by increasing the covalency of metal-oxygen bonds. However, Chunzhen Yang et al. note in Cation insertion to break the activity/stability relationship for highly active oxygen evolution reaction catalyst that these modifications frequently destabilize the catalyst surface, leading to rapid degradation.
This trade-off is particularly acute for non-noble metals. While transition metals like cobalt are cost-effective, they typically suffer from structural collapse or dissolution when exposed to acidic electrolytes. Even benchmark noble-metal catalysts are not immune; Jiandong Hu et al. point out in Synergistic Ruthenium-Doped Amorphous IrO x Matrix for Robust Oxygen Evolution that even iridium oxides (IrOx) face a limit where increasing activity compromises long-term structural integrity.
Mechanisms of Acidic Decay
Degradation in acidic media manifests through several distinct chemical pathways:
- Metal Leaching and Dissolution: In fuel cells, active metal centers can leach out of their support lattice. Kai Ao argues in Single iron site catalysts with increased metal-site loading via a high-temperature imprinting approach for proton exchange membrane fuel cells that this "demetallation" is a primary hurdle for replacing platinum with cheaper iron-based catalysts.
- Over-oxidation: Ruthenium is more active and cheaper than iridium, but it is notoriously unstable. Lu Li demonstrates in Lanthanide-regulating Ru-O covalency optimizes acidic oxygen evolution electrocatalysis that under high oxidative stress, ruthenium tends to transform into soluble species, leading to total catalyst loss.
- Chemical Attack by Byproducts: In Li-ion batteries, trace moisture reacts with electrolyte salts to form hydrofluoric acid (HF). Soo Kim et al. show in High-throughput computational design of cathode coatings for Li-ion batteries that this acid aggressively dissolves transition metals from the cathode, causing irreversible phase transformations.
Engineering Structural Integrity
To break the activity-stability deadlock, researchers are moving toward "designer" architectures that protect active sites without shielding them from the reaction.
One approach involves lattice engineering. Yong Feng et al. introduced a "breathing mode" in cobalt oxide by doping it with neodymium, which allows the catalyst to maintain structural integrity for over 800 hours in acid, as detailed in Breathing mode in Nd-CoOx for active and stable proton exchange membrane water electrolysis. Similarly, Qin Yin et al. used lithium-ion intercalation to dual-engineer the electronic structure and lattice strain of $RuO_2$, enhancing both its speed and its lifespan in RuO2 electronic structure and lattice strain dual engineering for enhanced acidic oxygen evolution reaction performance.
Another strategy uses sacrificial or protective layers. Chris Wolverton proposes the use of "HF-scavenging" coatings in First-Principles Thermodynamic Analysis of Hydrogen-fluoride Scavenging Cathode Coatings for Lithium-Ion Batteries, which neutralize harmful acids before they can reach the reactive cathode surface.
Accelerating Discovery with Active Learning
Because the design space for these materials is vast—involving millions of potential combinations of metals, dopants, and structures—researchers are increasingly turning to artificial intelligence. Raul A. Flores et al. developed an active learning framework in Active Learning Accelerated Discovery of Stable Iridium Oxide Polymorphs for the Oxygen Evolution Reaction that identified a new, globally stable phase of $IrO_3$ twice as fast as traditional methods.
By combining equivariant graph neural networks with quantum mechanical simulations, the authors of Active learning accelerated exploration of single-atom local environments in multimetallic systems for oxygen electrocatalysis, Hoje Chun et al., were able to navigate over 30,000 potential catalyst sites to find multimetallic configurations that reconcile the tension between high reaction activity and thermodynamic stability.
Key Foundational Work
High-throughput computational design of cathode coatings for Li-ion batteries (2016) — contributed 1 concept elements to this direction.
First-Principles Thermodynamic Analysis of Hydrogen-fluoride Scavenging Cathode Coatings for Lithium-Ion Batteries (2013) — contributed 1 concept elements to this direction.